⭐ 欢迎来到虫虫下载站! | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站

📄 memcmp.s

📁 用于嵌入式Linux系统的标准C的库函数
💻 S
字号:
/******************************************************************************* *  * Copyright (c) 1993 Intel Corporation *  * Intel hereby grants you permission to copy, modify, and distribute this * software and its documentation.  Intel grants this permission provided * that the above copyright notice appears in all copies and that both the * copyright notice and this permission notice appear in supporting * documentation.  In addition, Intel grants this permission provided that * you prominently mark as "not part of the original" any modifications * made to this software or documentation, and that the name of Intel * Corporation not be used in advertising or publicity pertaining to * distribution of the software or the documentation without specific, * written prior permission. *  * Intel Corporation provides this AS IS, WITHOUT ANY WARRANTY, EXPRESS OR * IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTY OF MERCHANTABILITY * OR FITNESS FOR A PARTICULAR PURPOSE.  Intel makes no guarantee or * representations regarding the use of, or the results of the use of, * the software and documentation in terms of correctness, accuracy, * reliability, currentness, or otherwise; and you rely on the software, * documentation and results solely at your own risk. * * IN NO EVENT SHALL INTEL BE LIABLE FOR ANY LOSS OF USE, LOSS OF BUSINESS, * LOSS OF PROFITS, INDIRECT, INCIDENTAL, SPECIAL OR CONSEQUENTIAL DAMAGES * OF ANY KIND.  IN NO EVENT SHALL INTEL'S TOTAL LIABILITY EXCEED THE SUM * PAID TO INTEL FOR THE PRODUCT LICENSED HEREUNDER. *  ******************************************************************************/	.file "memcmp.s"#ifdef	__PIC	.pic#endif#ifdef	__PID	.pid#endif/* * (c) copyright 1988,1993 Intel Corp., all rights reserved *//*	procedure memcmp  (optimized assembler version for the 80960K series)	result = memcmp (src1_addr, src2_addr, max_bytes)	compare the byte array pointed to by src1_addr to the byte array	pointed to by src2_addr.  Return 0 iff the arrays are equal, -1 iff 	src1_addr is lexicographically less than src2_addr, and 1 iff it is 	lexicographically greater.  Do not compare more than max_bytes bytes.	Undefined behavior will occur if the end of either source array	is in the last two words of the program's allocated memory space. 	This is so because memcmp fetches ahead.  Disallowing the fetch ahead 	would impose a severe performance penalty.	Strategy:	Fetch the source strings by words and compare the words until either 	a differing word is found or max_bytes is exhausted.  In the former 	case, move through the words to find the differing byte and return 	plus or minus one, appropriately.  In the latter case, return zero 	(equality).	Tactics:	1) Do NOT try to fetch the words in a word aligned manner because,	in my judgement, the performance degradation experienced due to 	non-aligned accesses does NOT outweigh the time and complexity added 	by the preamble that would be necessary to assure alignment.  This 	is supported by the intuition that most source arrays (even more 	true of most big source arrays) will be word aligned to begin with.	2) Rather than decrementing max_bytes to zero, I calculate the	address of the byte after the last byte of the source_1 array, and	quit when the source byte pointer passes that.  */	.globl _memcmp	.globl __memcmp	.leafproc _memcmp,__memcmp	.align 2_memcmp:#ifndef __PIC	lda	.Lrett,g14#else	lda	.Lrett-(.+8)(ip),g14#endif__memcmp:	mov	g14,g13		# preserve return address	ldconst 0,g14		# conform to register conventions	cmpibge	0,g2,Lequal_exit	# quit if max_bytes <= 0	addo	g0,g2,g2	# calculate byte addr of byte after last in src1.Lwloop:	cmpo	 g0,g2			ld	(g0), g5	# fetch word of source_1	bge 	 Lequal_exit	# quit (equal) if max_bytes exhausted	ld	(g1), g3	# fetch word of source_2	addo     4,g0,g0	# post-increment source_1 byte ptr	addo     4,g1,g1       	# post-increment source_2 byte ptr	cmpobe	 g5,g3,.Lwloop	# branch if source words are equal	ldconst 0xff,g4		# byte extraction mask	subo     4,g0,g0	# back up src1 pointer.Lcloop: and	g4,g5,g7  	# extract and compare individual bytes	and	g4,g3,g6	cmpobne	g7,g6,.diff	# branch if they are different	shlo	8,g4,g4		# position mask for next extraction	addo	1,g0,g0	cmpobl	g0,g2,.Lcloop	# quit if max_bytes is exhaustedLequal_exit:	mov	0,g0	bx	(g13).Lrett:		ret.diff:	bl	.neg		# arrays differ at current byte.  				/* return 1 or -1 appropriately */        mov     1,g0	bx	(g13).neg:	subi    1,0,g0.Lexit: 		bx	(g13)/* end or memcmp */

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -