Orange Pi5 kernel

^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   1) /* SPDX-License-Identifier: GPL-2.0-only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   3)  * Copyright (C) 2013 ARM Ltd.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   4)  * Copyright (C) 2013 Linaro.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   5)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   6)  * This code is based on glibc cortex strings work originally authored by Linaro
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   7)  * be found @
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   8)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   9)  * http://bazaar.launchpad.net/~linaro-toolchain-dev/cortex-strings/trunk/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  10)  * files/head:/src/aarch64/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  11)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  12) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  13) #include <linux/linkage.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  14) #include <asm/assembler.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  15) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  16) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  17)  * compare two strings
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  18)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  19)  * Parameters:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  20)  *  x0 - const string 1 pointer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  21)  *  x1 - const string 2 pointer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  22)  *  x2 - the maximal length to be compared
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  23)  * Returns:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  24)  *  x0 - an integer less than, equal to, or greater than zero if s1 is found,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  25)  *     respectively, to be less than, to match, or be greater than s2.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  26)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  27) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  28) #define REP8_01 0x0101010101010101
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  29) #define REP8_7f 0x7f7f7f7f7f7f7f7f
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  30) #define REP8_80 0x8080808080808080
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  31) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  32) /* Parameters and result.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  33) src1		.req	x0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  34) src2		.req	x1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  35) limit		.req	x2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  36) result		.req	x0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  37) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  38) /* Internal variables.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  39) data1		.req	x3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  40) data1w		.req	w3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  41) data2		.req	x4
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  42) data2w		.req	w4
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  43) has_nul		.req	x5
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  44) diff		.req	x6
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  45) syndrome	.req	x7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  46) tmp1		.req	x8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  47) tmp2		.req	x9
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  48) tmp3		.req	x10
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  49) zeroones	.req	x11
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  50) pos		.req	x12
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  51) limit_wd	.req	x13
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  52) mask		.req	x14
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  53) endloop		.req	x15
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  54) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  55) SYM_FUNC_START_WEAK_PI(strncmp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  56) 	cbz	limit, .Lret0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  57) 	eor	tmp1, src1, src2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  58) 	mov	zeroones, #REP8_01
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  59) 	tst	tmp1, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  60) 	b.ne	.Lmisaligned8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  61) 	ands	tmp1, src1, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  62) 	b.ne	.Lmutual_align
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  63) 	/* Calculate the number of full and partial words -1.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  64) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  65) 	* when limit is mulitply of 8, if not sub 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  66) 	* the judgement of last dword will wrong.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  67) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  68) 	sub	limit_wd, limit, #1 /* limit != 0, so no underflow.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  69) 	lsr	limit_wd, limit_wd, #3  /* Convert to Dwords.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  70) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  71) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  72) 	* NUL detection works on the principle that (X - 1) & (~X) & 0x80
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  73) 	* (=> (X - 1) & ~(X | 0x7f)) is non-zero iff a byte is zero, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  74) 	* can be done in parallel across the entire word.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  75) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  76) .Lloop_aligned:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  77) 	ldr	data1, [src1], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  78) 	ldr	data2, [src2], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  79) .Lstart_realigned:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  80) 	subs	limit_wd, limit_wd, #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  81) 	sub	tmp1, data1, zeroones
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  82) 	orr	tmp2, data1, #REP8_7f
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  83) 	eor	diff, data1, data2  /* Non-zero if differences found.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  84) 	csinv	endloop, diff, xzr, pl  /* Last Dword or differences.*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  85) 	bics	has_nul, tmp1, tmp2 /* Non-zero if NUL terminator.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  86) 	ccmp	endloop, #0, #0, eq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  87) 	b.eq	.Lloop_aligned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  88) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  89) 	/*Not reached the limit, must have found the end or a diff.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  90) 	tbz	limit_wd, #63, .Lnot_limit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  91) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  92) 	/* Limit % 8 == 0 => all bytes significant.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  93) 	ands	limit, limit, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  94) 	b.eq	.Lnot_limit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  95) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  96) 	lsl	limit, limit, #3    /* Bits -> bytes.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  97) 	mov	mask, #~0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  98) CPU_BE( lsr	mask, mask, limit )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  99) CPU_LE( lsl	mask, mask, limit )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) 	bic	data1, data1, mask
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) 	bic	data2, data2, mask
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) 	/* Make sure that the NUL byte is marked in the syndrome.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) 	orr	has_nul, has_nul, mask
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) .Lnot_limit:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) 	orr	syndrome, diff, has_nul
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) 	b	.Lcal_cmpresult
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) .Lmutual_align:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) 	* Sources are mutually aligned, but are not currently at an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) 	* alignment boundary.  Round down the addresses and then mask off
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) 	* the bytes that precede the start point.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) 	* We also need to adjust the limit calculations, but without
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) 	* overflowing if the limit is near ULONG_MAX.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) 	bic	src1, src1, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) 	bic	src2, src2, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) 	ldr	data1, [src1], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) 	neg	tmp3, tmp1, lsl #3  /* 64 - bits(bytes beyond align). */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) 	ldr	data2, [src2], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) 	mov	tmp2, #~0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) 	sub	limit_wd, limit, #1 /* limit != 0, so no underflow.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) 	/* Big-endian.  Early bytes are at MSB.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) CPU_BE( lsl	tmp2, tmp2, tmp3 )	/* Shift (tmp1 & 63).  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) 	/* Little-endian.  Early bytes are at LSB.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) CPU_LE( lsr	tmp2, tmp2, tmp3 )	/* Shift (tmp1 & 63).  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) 	and	tmp3, limit_wd, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) 	lsr	limit_wd, limit_wd, #3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) 	/* Adjust the limit. Only low 3 bits used, so overflow irrelevant.*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) 	add	limit, limit, tmp1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) 	add	tmp3, tmp3, tmp1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) 	orr	data1, data1, tmp2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) 	orr	data2, data2, tmp2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) 	add	limit_wd, limit_wd, tmp3, lsr #3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) 	b	.Lstart_realigned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) /*when src1 offset is not equal to src2 offset...*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) .Lmisaligned8:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) 	cmp	limit, #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) 	b.lo	.Ltiny8proc /*limit < 8... */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) 	* Get the align offset length to compare per byte first.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) 	* After this process, one string's address will be aligned.*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) 	and	tmp1, src1, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) 	neg	tmp1, tmp1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) 	add	tmp1, tmp1, #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) 	and	tmp2, src2, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) 	neg	tmp2, tmp2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) 	add	tmp2, tmp2, #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) 	subs	tmp3, tmp1, tmp2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) 	csel	pos, tmp1, tmp2, hi /*Choose the maximum. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) 	* Here, limit is not less than 8, so directly run .Ltinycmp
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) 	* without checking the limit.*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) 	sub	limit, limit, pos
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) .Ltinycmp:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) 	ldrb	data1w, [src1], #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) 	ldrb	data2w, [src2], #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) 	subs	pos, pos, #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) 	ccmp	data1w, #1, #0, ne  /* NZCV = 0b0000.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) 	ccmp	data1w, data2w, #0, cs  /* NZCV = 0b0000.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) 	b.eq	.Ltinycmp
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) 	cbnz	pos, 1f /*find the null or unequal...*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) 	cmp	data1w, #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) 	ccmp	data1w, data2w, #0, cs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) 	b.eq	.Lstart_align /*the last bytes are equal....*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) 1:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) 	sub	result, data1, data2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) 	ret
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) .Lstart_align:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) 	lsr	limit_wd, limit, #3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) 	cbz	limit_wd, .Lremain8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) 	/*process more leading bytes to make str1 aligned...*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) 	ands	xzr, src1, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) 	b.eq	.Lrecal_offset
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) 	add	src1, src1, tmp3	/*tmp3 is positive in this branch.*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) 	add	src2, src2, tmp3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) 	ldr	data1, [src1], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) 	ldr	data2, [src2], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) 	sub	limit, limit, tmp3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) 	lsr	limit_wd, limit, #3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) 	subs	limit_wd, limit_wd, #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) 	sub	tmp1, data1, zeroones
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) 	orr	tmp2, data1, #REP8_7f
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) 	eor	diff, data1, data2  /* Non-zero if differences found.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) 	csinv	endloop, diff, xzr, ne/*if limit_wd is 0,will finish the cmp*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) 	bics	has_nul, tmp1, tmp2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) 	ccmp	endloop, #0, #0, eq /*has_null is ZERO: no null byte*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) 	b.ne	.Lunequal_proc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) 	/*How far is the current str2 from the alignment boundary...*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) 	and	tmp3, tmp3, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) .Lrecal_offset:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) 	neg	pos, tmp3
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) .Lloopcmp_proc:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) 	* Divide the eight bytes into two parts. First,backwards the src2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) 	* to an alignment boundary,load eight bytes from the SRC2 alignment
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) 	* boundary,then compare with the relative bytes from SRC1.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) 	* If all 8 bytes are equal,then start the second part's comparison.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) 	* Otherwise finish the comparison.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) 	* This special handle can garantee all the accesses are in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) 	* thread/task space in avoid to overrange access.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) 	ldr	data1, [src1,pos]
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) 	ldr	data2, [src2,pos]
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) 	sub	tmp1, data1, zeroones
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) 	orr	tmp2, data1, #REP8_7f
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) 	bics	has_nul, tmp1, tmp2 /* Non-zero if NUL terminator.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) 	eor	diff, data1, data2  /* Non-zero if differences found.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) 	csinv	endloop, diff, xzr, eq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) 	cbnz	endloop, .Lunequal_proc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) 	/*The second part process*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) 	ldr	data1, [src1], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) 	ldr	data2, [src2], #8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) 	subs	limit_wd, limit_wd, #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) 	sub	tmp1, data1, zeroones
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) 	orr	tmp2, data1, #REP8_7f
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) 	eor	diff, data1, data2  /* Non-zero if differences found.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) 	csinv	endloop, diff, xzr, ne/*if limit_wd is 0,will finish the cmp*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) 	bics	has_nul, tmp1, tmp2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) 	ccmp	endloop, #0, #0, eq /*has_null is ZERO: no null byte*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) 	b.eq	.Lloopcmp_proc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) .Lunequal_proc:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) 	orr	syndrome, diff, has_nul
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) 	cbz	syndrome, .Lremain8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) .Lcal_cmpresult:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) 	* reversed the byte-order as big-endian,then CLZ can find the most
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) 	* significant zero bits.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) CPU_LE( rev	syndrome, syndrome )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) CPU_LE( rev	data1, data1 )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) CPU_LE( rev	data2, data2 )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) 	* For big-endian we cannot use the trick with the syndrome value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) 	* as carry-propagation can corrupt the upper bits if the trailing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) 	* bytes in the string contain 0x01.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) 	* However, if there is no NUL byte in the dword, we can generate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) 	* the result directly.  We can't just subtract the bytes as the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) 	* MSB might be significant.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) CPU_BE( cbnz	has_nul, 1f )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) CPU_BE( cmp	data1, data2 )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) CPU_BE( cset	result, ne )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) CPU_BE( cneg	result, result, lo )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) CPU_BE( ret )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) CPU_BE( 1: )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) 	/* Re-compute the NUL-byte detection, using a byte-reversed value.*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) CPU_BE( rev	tmp3, data1 )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) CPU_BE( sub	tmp1, tmp3, zeroones )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) CPU_BE( orr	tmp2, tmp3, #REP8_7f )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) CPU_BE( bic	has_nul, tmp1, tmp2 )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) CPU_BE( rev	has_nul, has_nul )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) CPU_BE( orr	syndrome, diff, has_nul )
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) 	* The MS-non-zero bit of the syndrome marks either the first bit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) 	* that is different, or the top bit of the first zero byte.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) 	* Shifting left now will bring the critical information into the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) 	* top bits.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) 	clz	pos, syndrome
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) 	lsl	data1, data1, pos
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) 	lsl	data2, data2, pos
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) 	* But we need to zero-extend (char is unsigned) the value and then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) 	* perform a signed 32-bit subtraction.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) 	*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) 	lsr	data1, data1, #56
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) 	sub	result, data1, data2, lsr #56
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) 	ret
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) .Lremain8:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) 	/* Limit % 8 == 0 => all bytes significant.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) 	ands	limit, limit, #7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) 	b.eq	.Lret0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) .Ltiny8proc:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) 	ldrb	data1w, [src1], #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) 	ldrb	data2w, [src2], #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) 	subs	limit, limit, #1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) 	ccmp	data1w, #1, #0, ne  /* NZCV = 0b0000.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) 	ccmp	data1w, data2w, #0, cs  /* NZCV = 0b0000.  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) 	b.eq	.Ltiny8proc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) 	sub	result, data1, data2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) 	ret
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) .Lret0:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) 	mov	result, #0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) 	ret
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) SYM_FUNC_END_PI(strncmp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) EXPORT_SYMBOL_NOKASAN(strncmp)