| Austin Schuh | dace2a6 | 2020-08-18 10:56:48 -0700 | [diff] [blame] | 1 | dnl x86 mpn_rshift -- mpn right shift. |
| 2 | |
| 3 | dnl Copyright 1992, 1994, 1996, 1999-2002 Free Software Foundation, Inc. |
| 4 | |
| 5 | dnl This file is part of the GNU MP Library. |
| 6 | dnl |
| 7 | dnl The GNU MP Library is free software; you can redistribute it and/or modify |
| 8 | dnl it under the terms of either: |
| 9 | dnl |
| 10 | dnl * the GNU Lesser General Public License as published by the Free |
| 11 | dnl Software Foundation; either version 3 of the License, or (at your |
| 12 | dnl option) any later version. |
| 13 | dnl |
| 14 | dnl or |
| 15 | dnl |
| 16 | dnl * the GNU General Public License as published by the Free Software |
| 17 | dnl Foundation; either version 2 of the License, or (at your option) any |
| 18 | dnl later version. |
| 19 | dnl |
| 20 | dnl or both in parallel, as here. |
| 21 | dnl |
| 22 | dnl The GNU MP Library is distributed in the hope that it will be useful, but |
| 23 | dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |
| 24 | dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
| 25 | dnl for more details. |
| 26 | dnl |
| 27 | dnl You should have received copies of the GNU General Public License and the |
| 28 | dnl GNU Lesser General Public License along with the GNU MP Library. If not, |
| 29 | dnl see https://www.gnu.org/licenses/. |
| 30 | |
| 31 | include(`../config.m4') |
| 32 | |
| 33 | |
| 34 | C cycles/limb |
| 35 | C P54 7.5 |
| 36 | C P55 7.0 |
| 37 | C P6 2.5 |
| 38 | C K6 4.5 |
| 39 | C K7 5.0 |
| 40 | C P4 16.5 |
| 41 | |
| 42 | |
| 43 | C mp_limb_t mpn_rshift (mp_ptr dst, mp_srcptr src, mp_size_t size, |
| 44 | C unsigned shift); |
| 45 | |
| 46 | defframe(PARAM_SHIFT,16) |
| 47 | defframe(PARAM_SIZE, 12) |
| 48 | defframe(PARAM_SRC, 8) |
| 49 | defframe(PARAM_DST, 4) |
| 50 | |
| 51 | TEXT |
| 52 | ALIGN(8) |
| 53 | PROLOGUE(mpn_rshift) |
| 54 | |
| 55 | pushl %edi |
| 56 | pushl %esi |
| 57 | pushl %ebx |
| 58 | deflit(`FRAME',12) |
| 59 | |
| 60 | movl PARAM_DST,%edi |
| 61 | movl PARAM_SRC,%esi |
| 62 | movl PARAM_SIZE,%edx |
| 63 | movl PARAM_SHIFT,%ecx |
| 64 | |
| 65 | leal -4(%edi,%edx,4),%edi |
| 66 | leal (%esi,%edx,4),%esi |
| 67 | negl %edx |
| 68 | |
| 69 | movl (%esi,%edx,4),%ebx C read least significant limb |
| 70 | xorl %eax,%eax |
| 71 | shrdl( %cl, %ebx, %eax) C compute carry limb |
| 72 | incl %edx |
| 73 | jz L(end) |
| 74 | pushl %eax C push carry limb onto stack |
| 75 | testb $1,%dl |
| 76 | jnz L(1) C enter loop in the middle |
| 77 | movl %ebx,%eax |
| 78 | |
| 79 | ALIGN(8) |
| 80 | L(oop): movl (%esi,%edx,4),%ebx C load next higher limb |
| 81 | shrdl( %cl, %ebx, %eax) C compute result limb |
| 82 | movl %eax,(%edi,%edx,4) C store it |
| 83 | incl %edx |
| 84 | L(1): movl (%esi,%edx,4),%eax |
| 85 | shrdl( %cl, %eax, %ebx) |
| 86 | movl %ebx,(%edi,%edx,4) |
| 87 | incl %edx |
| 88 | jnz L(oop) |
| 89 | |
| 90 | shrl %cl,%eax C compute most significant limb |
| 91 | movl %eax,(%edi) C store it |
| 92 | |
| 93 | popl %eax C pop carry limb |
| 94 | |
| 95 | popl %ebx |
| 96 | popl %esi |
| 97 | popl %edi |
| 98 | ret |
| 99 | |
| 100 | L(end): shrl %cl,%ebx C compute most significant limb |
| 101 | movl %ebx,(%edi) C store it |
| 102 | |
| 103 | popl %ebx |
| 104 | popl %esi |
| 105 | popl %edi |
| 106 | ret |
| 107 | |
| 108 | EPILOGUE() |