1/* 2 Copyright (c) 2015-2024, Synopsys, Inc. All rights reserved. 3 4 Redistribution and use in source and binary forms, with or without 5 modification, are permitted provided that the following conditions are met: 6 7 1) Redistributions of source code must retain the above copyright notice, 8 this list of conditions and the following disclaimer. 9 10 2) Redistributions in binary form must reproduce the above copyright notice, 11 this list of conditions and the following disclaimer in the documentation 12 and/or other materials provided with the distribution. 13 14 3) Neither the name of the Synopsys, Inc., nor the names of its contributors 15 may be used to endorse or promote products derived from this software 16 without specific prior written permission. 17 18 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 19 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE 22 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 23 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 24 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 25 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 26 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 27 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 28 POSSIBILITY OF SUCH DAMAGE. 29*/ 30 31/* This implementation is optimized for performance. For code size a generic 32 implementation of this function from newlib/libc/string/strcmp.c will be 33 used. */ 34#include <picolibc.h> 35 36#if !defined (__OPTIMIZE_SIZE__) && !defined (PREFER_SIZE_OVER_SPEED) \ 37 && !defined (__ARC_RF16__) 38 39#include "asm.h" 40 41/* This is optimized primarily for the ARC700. 42 It would be possible to speed up the loops by one cycle / word 43 respective one cycle / byte by forcing double source 1 alignment, unrolling 44 by a factor of two, and speculatively loading the second word / byte of 45 source 1; however, that would increase the overhead for loop setup / finish, 46 and strcmp might often terminate early. */ 47#ifndef __ARCHS__ 48 49ENTRY (strcmp) 50 or r2,r0,r1 51 bmsk_s r2,r2,1 52 brne_l r2,0,.Lcharloop 53 mov_s r12,0x01010101 54 ror r5,r12 55.Lwordloop: 56 ld.ab r2,[r0,4] 57 ld.ab r3,[r1,4] 58 nop_s 59 sub r4,r2,r12 60 bic r4,r4,r2 61 and r4,r4,r5 62 brne_l r4,0,.Lfound0 63 breq r2,r3,.Lwordloop 64#ifdef __LITTLE_ENDIAN__ 65 xor r0,r2,r3 ; mask for difference 66 sub_s r1,r0,1 67 bic_s r0,r0,r1 ; mask for least significant difference bit 68 sub r1,r5,r0 69 xor r0,r5,r1 ; mask for least significant difference byte 70 and_s r2,r2,r0 71 and_s r3,r3,r0 72#endif /* LITTLE ENDIAN */ 73 cmp_s r2,r3 74 mov_s r0,1 75 j_s.d [blink] 76 bset.lo r0,r0,31 77 78 .balign 4 79#ifdef __LITTLE_ENDIAN__ 80.Lfound0: 81 xor r0,r2,r3 ; mask for difference 82 or r0,r0,r4 ; or in zero indicator 83 sub_s r1,r0,1 84 bic_s r0,r0,r1 ; mask for least significant difference bit 85 sub r1,r5,r0 86 xor r0,r5,r1 ; mask for least significant difference byte 87 and_s r2,r2,r0 88 and_s r3,r3,r0 89 sub.f r0,r2,r3 90 mov.hi r0,1 91 j_s.d [blink] 92 bset.lo r0,r0,31 93#else /* BIG ENDIAN */ 94 /* The zero-detection above can mis-detect 0x01 bytes as zeroes 95 because of carry-propagateion from a lower significant zero byte. 96 We can compensate for this by checking that bit0 is zero. 97 This compensation is not necessary in the step where we 98 get a low estimate for r2, because in any affected bytes 99 we already have 0x00 or 0x01, which will remain unchanged 100 when bit 7 is cleared. */ 101 .balign 4 102.Lfound0: 103#ifdef __ARC_BARREL_SHIFTER__ 104 lsr r0,r4,8 105 lsr_s r1,r2 106 bic_s r2,r2,r0 ; get low estimate for r2 and get ... 107 bic_s r0,r0,r1 ; <this is the adjusted mask for zeros> 108 or_s r3,r3,r0 ; ... high estimate r3 so that r2 > r3 will ... 109 cmp_s r3,r2 ; ... be independent of trailing garbage 110 or_s r2,r2,r0 ; likewise for r3 > r2 111 bic_s r3,r3,r0 112 rlc r0,0 ; r0 := r2 > r3 ? 1 : 0 113 cmp_s r2,r3 114 j_s.d [blink] 115 bset.lo r0,r0,31 116#else /* __ARC_BARREL_SHIFTER__ */ 117 /* Fall through to .Lcharloop. */ 118 sub_s r0,r0,4 119 sub_s r1,r1,4 120#endif /* __ARC_BARREL_SHIFTER__ */ 121#endif /* ENDIAN */ 122 123 .balign 4 124.Lcharloop: 125 ldb.ab r2,[r0,1] 126 ldb.ab r3,[r1,1] 127 nop_s 128 breq_l r2,0,.Lcmpend 129 breq r2,r3,.Lcharloop 130.Lcmpend: 131 j_s.d [blink] 132 sub r0,r2,r3 133ENDFUNC (strcmp) 134#endif /* !__ARCHS__ */ 135 136#endif /* !__OPTIMIZE_SIZE__ && !PREFER_SIZE_OVER_SPEED */ 137
