1dnl ARM mpn_bdiv_dbm1c. 2 3dnl Copyright 2008, 2011, 2012 Free Software Foundation, Inc. 4 5dnl This file is part of the GNU MP Library. 6dnl 7dnl The GNU MP Library is free software; you can redistribute it and/or modify 8dnl it under the terms of either: 9dnl 10dnl * the GNU Lesser General Public License as published by the Free 11dnl Software Foundation; either version 3 of the License, or (at your 12dnl option) any later version. 13dnl 14dnl or 15dnl 16dnl * the GNU General Public License as published by the Free Software 17dnl Foundation; either version 2 of the License, or (at your option) any 18dnl later version. 19dnl 20dnl or both in parallel, as here. 21dnl 22dnl The GNU MP Library is distributed in the hope that it will be useful, but 23dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY 24dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 25dnl for more details. 26dnl 27dnl You should have received copies of the GNU General Public License and the 28dnl GNU Lesser General Public License along with the GNU MP Library. If not, 29dnl see https://www.gnu.org/licenses/. 30 31include(`../config.m4') 32 33C cycles/limb 34C StrongARM ? 35C XScale ? 36C Cortex-A7 ? 37C Cortex-A8 ? 38C Cortex-A9 4.25 39C Cortex-A15 2.5 40 41C TODO 42C * Try using umlal or umaal. 43C * Try using ldm/stm. 44 45define(`qp', `r0') 46define(`up', `r1') 47define(`n', `r2') 48define(`bd', `r3') 49define(`cy', `sp,#0') 50 51ASM_START() 52 TEXT 53 ALIGN(16) 54PROLOGUE(mpn_bdiv_dbm1c) 55 push {r4, r5, r6, r7, r8} 56 ldr r4, [up], #4 57 ldr r5, [sp, #20] 58 ands r12, n, #3 59 beq L(fi0) 60 cmp r12, #2 61 bcc L(fi1) 62 beq L(fi2) 63 64L(fi3): umull r8, r12, r4, bd 65 ldr r4, [up], #4 66 b L(lo3) 67 68L(fi0): umull r6, r7, r4, bd 69 ldr r4, [up], #4 70 b L(lo0) 71 72L(fi1): subs n, n, #1 73 umull r8, r12, r4, bd 74 bls L(wd1) 75 ldr r4, [up], #4 76 b L(lo1) 77 78L(fi2): umull r6, r7, r4, bd 79 ldr r4, [up], #4 80 b L(lo2) 81 82L(top): ldr r4, [up], #4 83 subs r5, r5, r6 84 str r5, [qp], #4 85 sbc r5, r5, r7 86L(lo1): umull r6, r7, r4, bd 87 ldr r4, [up], #4 88 subs r5, r5, r8 89 str r5, [qp], #4 90 sbc r5, r5, r12 91L(lo0): umull r8, r12, r4, bd 92 ldr r4, [up], #4 93 subs r5, r5, r6 94 str r5, [qp], #4 95 sbc r5, r5, r7 96L(lo3): umull r6, r7, r4, bd 97 ldr r4, [up], #4 98 subs r5, r5, r8 99 str r5, [qp], #4 100 sbc r5, r5, r12 101L(lo2): subs n, n, #4 102 umull r8, r12, r4, bd 103 bhi L(top) 104 105L(wd2): subs r5, r5, r6 106 str r5, [qp], #4 107 sbc r5, r5, r7 108L(wd1): subs r5, r5, r8 109 str r5, [qp] 110 sbc r0, r5, r12 111 pop {r4, r5, r6, r7, r8} 112 return lr 113EPILOGUE() 114