projects / deliverable / linux.git / blame
| Commit | Line | Data |
|---|---|---|
| b920de1b DH |
1 | /* Optimised simple memory fill |
| 2 | * | |
| 3 | * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved. | |
| 4 | * Written by David Howells (dhowells@redhat.com) | |
| 5 | * | |
| 6 | * This program is free software; you can redistribute it and/or | |
| 7 | * modify it under the terms of the GNU General Public Licence | |
| 8 | * as published by the Free Software Foundation; either version | |
| 9 | * 2 of the Licence, or (at your option) any later version. | |
| 10 | */ | |
| 11 | #include <asm/cache.h> | |
| 12 | ||
| 13 | .section .text | |
| 14 | .balign L1_CACHE_BYTES | |
| 15 | ||
| 16 | ############################################################################### | |
| 17 | # | |
| 18 | # void *memset(void *dst, int c, size_t n) | |
| 19 | # | |
| 20 | ############################################################################### | |
| 21 | .globl memset | |
| 22 | .type memset,@function | |
| 23 | memset: | |
| 24 | movm [d2,d3],(sp) | |
| 25 | mov d0,(12,sp) | |
| 26 | mov d1,(16,sp) | |
| 27 | mov (20,sp),d2 # count | |
| 28 | mov d0,a0 # dst | |
| 29 | mov d0,e3 # the return value | |
| 30 | ||
| 31 | cmp +0,d2 | |
| 32 | beq memset_done # return if zero-length fill | |
| 33 | ||
| 34 | # see if the region parameters are four-byte aligned | |
| 35 | or d0,d2,d3 | |
| 36 | and +3,d3 | |
| 37 | bne memset_1 # jump if not | |
| 38 | ||
| 39 | extbu d1 | |
| 40 | mov_asl d1,d3,8,d1 | |
| 41 | or_asl d1,d3,8,d1 | |
| 42 | or_asl d1,d3,8,d1 | |
| 43 | or d3,d1 | |
| 44 | ||
| 45 | # we want to transfer as much as we can in chunks of 32 bytes | |
| 46 | cmp +31,d2 | |
| 47 | bls memset_4_remainder # 4-byte aligned remainder | |
| 48 | ||
| 49 | add -32,d2 | |
| 50 | mov +32,d3 | |
| 51 | ||
| 52 | memset_4_loop: | |
| 53 | mov d1,(a0+) | |
| 54 | mov d1,(a0+) | |
| 55 | mov d1,(a0+) | |
| 56 | mov d1,(a0+) | |
| 57 | mov d1,(a0+) | |
| 58 | mov d1,(a0+) | |
| 59 | mov d1,(a0+) | |
| 60 | mov d1,(a0+) | |
| 61 | ||
| 62 | sub d3,d2 | |
| 63 | bcc memset_4_loop | |
| 64 | ||
| 65 | add d3,d2 | |
| 66 | beq memset_4_no_remainder | |
| 67 | ||
| 68 | memset_4_remainder: | |
| 69 | # cut 4-7 words down to 0-3 | |
| 70 | cmp +16,d2 | |
| 71 | bcs memset_4_three_or_fewer_words | |
| 72 | mov d1,(a0+) | |
| 73 | mov d1,(a0+) | |
| 74 | mov d1,(a0+) | |
| 75 | mov d1,(a0+) | |
| 76 | add -16,d2 | |
| 77 | beq memset_4_no_remainder | |
| 78 | ||
| 79 | # copy the remaining 1, 2 or 3 words | |
| 80 | memset_4_three_or_fewer_words: | |
| 81 | cmp +8,d2 | |
| 82 | bcs memset_4_one_word | |
| 83 | beq memset_4_two_words | |
| 84 | mov d1,(a0+) | |
| 85 | memset_4_two_words: | |
| 86 | mov d1,(a0+) | |
| 87 | memset_4_one_word: | |
| 88 | mov d1,(a0+) | |
| 89 | ||
| 90 | memset_4_no_remainder: | |
| 91 | # check we set the correct amount | |
| 92 | # TODO: REMOVE CHECK | |
| 93 | sub e3,a0,d2 | |
| 94 | mov (20,sp),d1 | |
| 95 | cmp d2,d1 | |
| 96 | beq memset_done | |
| 97 | break | |
| 98 | break | |
| 99 | break | |
| 100 | ||
| 101 | memset_done: | |
| 102 | mov e3,a0 | |
| 103 | ret [d2,d3],8 | |
| 104 | ||
| 105 | # handle misaligned copying | |
| 106 | memset_1: | |
| 107 | add -1,d2 | |
| 108 | mov +1,d3 | |
| 109 | setlb # setlb requires the next insns | |
| 110 | # to occupy exactly 4 bytes | |
| 111 | ||
| 112 | sub d3,d2 | |
| 113 | movbu d1,(a0) | |
| 114 | inc a0 | |
| 115 | lcc | |
| 116 | ||
| 117 | mov e3,a0 | |
| 118 | ret [d2,d3],8 | |
| 119 | ||
| 120 | memset_end: | |
| 121 | .size memset, memset_end-memset |