| 55e84dc6 |
1 | /* Optimized memrchr implementation using LoongArch LASX instructions. |
| dff8da6b |
2 | Copyright (C) 2023-2024 Free Software Foundation, Inc. |
| 55e84dc6 |
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 <sysdep.h> |
| 20 | #include <sys/regdef.h> |
| 21 | #include <sys/asm.h> |
| 22 | |
| 23 | #if IS_IN (libc) && !defined __loongarch_soft_float |
| 24 | |
| 25 | #ifndef MEMRCHR |
| 26 | # define MEMRCHR __memrchr_lasx |
| 27 | #endif |
| 28 | |
| 29 | LEAF(MEMRCHR, 6) |
| 30 | beqz a2, L(ret0) |
| 31 | addi.d a2, a2, -1 |
| 32 | add.d a3, a0, a2 |
| 33 | andi t1, a3, 0x3f |
| 34 | |
| 35 | bstrins.d a3, zero, 5, 0 |
| 36 | addi.d t1, t1, 1 |
| 37 | xvld xr0, a3, 0 |
| 38 | xvld xr1, a3, 32 |
| 39 | |
| 40 | sub.d t2, zero, t1 |
| 41 | li.d t3, -1 |
| 42 | xvreplgr2vr.b xr2, a1 |
| 43 | andi t4, a0, 0x3f |
| 44 | |
| 45 | srl.d t2, t3, t2 |
| 46 | xvseq.b xr0, xr0, xr2 |
| 47 | xvseq.b xr1, xr1, xr2 |
| 48 | xvmsknz.b xr0, xr0 |
| 49 | |
| 50 | |
| 51 | xvmsknz.b xr1, xr1 |
| 52 | xvpickve.w xr3, xr0, 4 |
| 53 | xvpickve.w xr4, xr1, 4 |
| 54 | vilvl.h vr0, vr3, vr0 |
| 55 | |
| 56 | vilvl.h vr1, vr4, vr1 |
| 57 | vilvl.w vr0, vr1, vr0 |
| 58 | movfr2gr.d t0, fa0 |
| 59 | and t0, t0, t2 |
| 60 | |
| 61 | bltu a2, t1, L(end) |
| 62 | bnez t0, L(found) |
| 63 | bstrins.d a0, zero, 5, 0 |
| 64 | L(loop): |
| 65 | xvld xr0, a3, -64 |
| 66 | |
| 67 | xvld xr1, a3, -32 |
| 68 | addi.d a3, a3, -64 |
| 69 | xvseq.b xr0, xr0, xr2 |
| 70 | xvseq.b xr1, xr1, xr2 |
| 71 | |
| 72 | |
| 73 | beq a0, a3, L(out) |
| 74 | xvmax.bu xr3, xr0, xr1 |
| 75 | xvseteqz.v fcc0, xr3 |
| 76 | bcnez fcc0, L(loop) |
| 77 | |
| 78 | xvmsknz.b xr0, xr0 |
| 79 | xvmsknz.b xr1, xr1 |
| 80 | xvpickve.w xr3, xr0, 4 |
| 81 | xvpickve.w xr4, xr1, 4 |
| 82 | |
| 83 | vilvl.h vr0, vr3, vr0 |
| 84 | vilvl.h vr1, vr4, vr1 |
| 85 | vilvl.w vr0, vr1, vr0 |
| 86 | movfr2gr.d t0, fa0 |
| 87 | |
| 88 | L(found): |
| 89 | addi.d a0, a3, 63 |
| 90 | clz.d t1, t0 |
| 91 | sub.d a0, a0, t1 |
| 92 | jr ra |
| 93 | |
| 94 | |
| 95 | L(out): |
| 96 | xvmsknz.b xr0, xr0 |
| 97 | xvmsknz.b xr1, xr1 |
| 98 | xvpickve.w xr3, xr0, 4 |
| 99 | xvpickve.w xr4, xr1, 4 |
| 100 | |
| 101 | vilvl.h vr0, vr3, vr0 |
| 102 | vilvl.h vr1, vr4, vr1 |
| 103 | vilvl.w vr0, vr1, vr0 |
| 104 | movfr2gr.d t0, fa0 |
| 105 | |
| 106 | L(end): |
| 107 | sll.d t2, t3, t4 |
| 108 | and t0, t0, t2 |
| 109 | addi.d a0, a3, 63 |
| 110 | clz.d t1, t0 |
| 111 | |
| 112 | sub.d a0, a0, t1 |
| 113 | maskeqz a0, a0, t0 |
| 114 | jr ra |
| 115 | L(ret0): |
| 116 | move a0, zero |
| 117 | |
| 118 | |
| 119 | jr ra |
| 120 | END(MEMRCHR) |
| 121 | |
| 122 | libc_hidden_builtin_def (MEMRCHR) |
| 123 | #endif |
projects / thirdparty / glibc.git / blame