summaryrefslogtreecommitdiffstats
path: root/arch/hexagon/mm/strnlen_user.S
blob: 5c6a16c7c72ac9527e6858fd82f17428524fba5b (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
/*
 * User string length functions for kernel
 *
 * Copyright (c) 2010-2011, Code Aurora Forum. All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 and
 * only version 2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
 * 02110-1301, USA.
 */

#define isrc	r0
#define max	r1	/*  Do not change!  */

#define end	r2
#define tmp1	r3

#define obo	r6	/*  off-by-one  */
#define start	r7
#define mod8	r8
#define dbuf    r15:14
#define dcmp	r13:12

/*
 * The vector mask version of this turned out *really* badly.
 * The hardware loop version also turned out *really* badly.
 * Seems straight pointer arithmetic basically wins here.
 */

#define fname __strnlen_user

	.text
	.global fname
	.type fname, @function
	.p2align 5  /*  why?  */
fname:
	{
		mod8 = and(isrc,#7);
		end = add(isrc,max);
		start = isrc;
	}
	{
		P0 = cmp.eq(mod8,#0);
		mod8 = and(end,#7);
		dcmp = #0;
		if (P0.new) jump:t dw_loop;	/*  fire up the oven  */
	}

alignment_loop:
fail_1:	{
		tmp1 = memb(start++#1);
	}
	{
		P0 = cmp.eq(tmp1,#0);
		if (P0.new) jump:nt exit_found;
		P1 = cmp.gtu(end,start);
		mod8 = and(start,#7);
	}
	{
		if (!P1) jump exit_error;  /*  hit the end  */
		P0 = cmp.eq(mod8,#0);
	}
	{
		if (!P0) jump alignment_loop;
	}



dw_loop:
fail_2:	{
		dbuf = memd(start);
		obo = add(start,#1);
	}
	{
		P0 = vcmpb.eq(dbuf,dcmp);
	}
	{
		tmp1 = P0;
		P0 = cmp.gtu(end,start);
	}
	{
		tmp1 = ct0(tmp1);
		mod8 = and(end,#7);
		if (!P0) jump end_check;
	}
	{
		P0 = cmp.eq(tmp1,#32);
		if (!P0.new) jump:nt exit_found;
		if (!P0.new) start = add(obo,tmp1);
	}
	{
		start = add(start,#8);
		jump dw_loop;
	}	/*  might be nice to combine these jumps...   */


end_check:
	{
		P0 = cmp.gt(tmp1,mod8);
		if (P0.new) jump:nt exit_error;	/*  neverfound!  */
		start = add(obo,tmp1);
	}

exit_found:
	{
		R0 = sub(start,isrc);
		jumpr R31;
	}

exit_error:
	{
		R0 = add(max,#1);
		jumpr R31;
	}

	/*  Uh, what does the "fixup" return here?  */
	.falign
fix_1:
	{
		R0 = #0;
		jumpr R31;
	}

	.size fname,.-fname


.section __ex_table,"a"
.long fail_1,fix_1
.long fail_2,fix_1
.previous
OpenPOWER on IntegriCloud