blob: 8422f38e12183561855648e6c14fd8d042203185 [file] [log] [blame]
Vineet Gupta5210d1e2013-01-18 15:12:18 +05301/*
2 * Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License version 2 as
6 * published by the Free Software Foundation.
7 */
8
9/* If dst and src are 4 byte aligned, copy 8 bytes at a time.
10 If the src is 4, but not 8 byte aligned, we first read 4 bytes to get
11 it 8 byte aligned. Thus, we can do a little read-ahead, without
12 dereferencing a cache line that we should not touch.
13 Note that short and long instructions have been scheduled to avoid
14 branch stalls.
15 The beq_s to r3z could be made unaligned & long to avoid a stall
16 there, but the it is not likely to be taken often, and it
17 would also be likey to cost an unaligned mispredict at the next call. */
18
Vineet Guptaec7ac6a2014-02-07 13:47:43 +053019#include <linux/linkage.h>
Vineet Gupta5210d1e2013-01-18 15:12:18 +053020
Vineet Guptaec7ac6a2014-02-07 13:47:43 +053021ENTRY(strcpy)
Vineet Gupta5210d1e2013-01-18 15:12:18 +053022 or r2,r0,r1
23 bmsk_s r2,r2,1
24 brne.d r2,0,charloop
25 mov_s r10,r0
26 ld_s r3,[r1,0]
27 mov r8,0x01010101
28 bbit0.d r1,2,loop_start
29 ror r12,r8
30 sub r2,r3,r8
31 bic_s r2,r2,r3
32 tst_s r2,r12
33 bne r3z
34 mov_s r4,r3
35 .balign 4
36loop:
37 ld.a r3,[r1,4]
38 st.ab r4,[r10,4]
39loop_start:
40 ld.a r4,[r1,4]
41 sub r2,r3,r8
42 bic_s r2,r2,r3
43 tst_s r2,r12
44 bne_s r3z
45 st.ab r3,[r10,4]
46 sub r2,r4,r8
47 bic r2,r2,r4
48 tst r2,r12
49 beq loop
50 mov_s r3,r4
51#ifdef __LITTLE_ENDIAN__
52r3z: bmsk.f r1,r3,7
53 lsr_s r3,r3,8
54#else
55r3z: lsr.f r1,r3,24
56 asl_s r3,r3,8
57#endif
58 bne.d r3z
59 stb.ab r1,[r10,1]
60 j_s [blink]
61
62 .balign 4
63charloop:
64 ldb.ab r3,[r1,1]
65
66
67 brne.d r3,0,charloop
68 stb.ab r3,[r10,1]
69 j [blink]
Vineet Guptaec7ac6a2014-02-07 13:47:43 +053070END(strcpy)