1 | /* SPDX-License-Identifier: GPL-2.0+ WITH GCC-exception-2.0 |
2 | |
3 | Copyright (C) 1994, 1995, 1997, 1998, 1999, 2000, 2001, 2002, 2003, |
4 | 2004, 2005, 2006 |
5 | Free Software Foundation, Inc. |
6 | */ |
7 | |
8 | !! libgcc routines for the Renesas / SuperH SH CPUs. |
9 | !! Contributed by Steve Chamberlain. |
10 | !! sac@cygnus.com |
11 | |
12 | !! ashiftrt_r4_x, ___ashrsi3, ___ashlsi3, ___lshrsi3 routines |
13 | !! recoded in assembly by Toshiyasu Morita |
14 | !! tm@netcom.com |
15 | |
16 | /* SH2 optimizations for ___ashrsi3, ___ashlsi3, ___lshrsi3 and |
17 | ELF local label prefixes by J"orn Rennecke |
18 | amylaar@cygnus.com */ |
19 | |
20 | .text |
21 | .balign 4 |
22 | .global __movmem |
23 | .global __movstr |
24 | .set __movstr, __movmem |
25 | /* This would be a lot simpler if r6 contained the byte count |
26 | minus 64, and we wouldn't be called here for a byte count of 64. */ |
27 | __movmem: |
28 | sts.l pr,@-r15 |
29 | shll2 r6 |
30 | bsr __movmemSI52+2 |
31 | mov.l @(48,r5),r0 |
32 | .balign 4 |
33 | movmem_loop: /* Reached with rts */ |
34 | mov.l @(60,r5),r0 |
35 | add #-64,r6 |
36 | mov.l r0,@(60,r4) |
37 | tst r6,r6 |
38 | mov.l @(56,r5),r0 |
39 | bt movmem_done |
40 | mov.l r0,@(56,r4) |
41 | cmp/pl r6 |
42 | mov.l @(52,r5),r0 |
43 | add #64,r5 |
44 | mov.l r0,@(52,r4) |
45 | add #64,r4 |
46 | bt __movmemSI52 |
47 | ! done all the large groups, do the remainder |
48 | ! jump to movmem+ |
49 | mova __movmemSI4+4,r0 |
50 | add r6,r0 |
51 | jmp @r0 |
52 | movmem_done: ! share slot insn, works out aligned. |
53 | lds.l @r15+,pr |
54 | mov.l r0,@(56,r4) |
55 | mov.l @(52,r5),r0 |
56 | rts |
57 | mov.l r0,@(52,r4) |
58 | .balign 4 |
59 | |
60 | .global __movmemSI64 |
61 | .global __movstrSI64 |
62 | .set __movstrSI64, __movmemSI64 |
63 | __movmemSI64: |
64 | mov.l @(60,r5),r0 |
65 | mov.l r0,@(60,r4) |
66 | .global __movmemSI60 |
67 | .global __movstrSI60 |
68 | .set __movstrSI60, __movmemSI60 |
69 | __movmemSI60: |
70 | mov.l @(56,r5),r0 |
71 | mov.l r0,@(56,r4) |
72 | .global __movmemSI56 |
73 | .global __movstrSI56 |
74 | .set __movstrSI56, __movmemSI56 |
75 | __movmemSI56: |
76 | mov.l @(52,r5),r0 |
77 | mov.l r0,@(52,r4) |
78 | .global __movmemSI52 |
79 | .global __movstrSI52 |
80 | .set __movstrSI52, __movmemSI52 |
81 | __movmemSI52: |
82 | mov.l @(48,r5),r0 |
83 | mov.l r0,@(48,r4) |
84 | .global __movmemSI48 |
85 | .global __movstrSI48 |
86 | .set __movstrSI48, __movmemSI48 |
87 | __movmemSI48: |
88 | mov.l @(44,r5),r0 |
89 | mov.l r0,@(44,r4) |
90 | .global __movmemSI44 |
91 | .global __movstrSI44 |
92 | .set __movstrSI44, __movmemSI44 |
93 | __movmemSI44: |
94 | mov.l @(40,r5),r0 |
95 | mov.l r0,@(40,r4) |
96 | .global __movmemSI40 |
97 | .global __movstrSI40 |
98 | .set __movstrSI40, __movmemSI40 |
99 | __movmemSI40: |
100 | mov.l @(36,r5),r0 |
101 | mov.l r0,@(36,r4) |
102 | .global __movmemSI36 |
103 | .global __movstrSI36 |
104 | .set __movstrSI36, __movmemSI36 |
105 | __movmemSI36: |
106 | mov.l @(32,r5),r0 |
107 | mov.l r0,@(32,r4) |
108 | .global __movmemSI32 |
109 | .global __movstrSI32 |
110 | .set __movstrSI32, __movmemSI32 |
111 | __movmemSI32: |
112 | mov.l @(28,r5),r0 |
113 | mov.l r0,@(28,r4) |
114 | .global __movmemSI28 |
115 | .global __movstrSI28 |
116 | .set __movstrSI28, __movmemSI28 |
117 | __movmemSI28: |
118 | mov.l @(24,r5),r0 |
119 | mov.l r0,@(24,r4) |
120 | .global __movmemSI24 |
121 | .global __movstrSI24 |
122 | .set __movstrSI24, __movmemSI24 |
123 | __movmemSI24: |
124 | mov.l @(20,r5),r0 |
125 | mov.l r0,@(20,r4) |
126 | .global __movmemSI20 |
127 | .global __movstrSI20 |
128 | .set __movstrSI20, __movmemSI20 |
129 | __movmemSI20: |
130 | mov.l @(16,r5),r0 |
131 | mov.l r0,@(16,r4) |
132 | .global __movmemSI16 |
133 | .global __movstrSI16 |
134 | .set __movstrSI16, __movmemSI16 |
135 | __movmemSI16: |
136 | mov.l @(12,r5),r0 |
137 | mov.l r0,@(12,r4) |
138 | .global __movmemSI12 |
139 | .global __movstrSI12 |
140 | .set __movstrSI12, __movmemSI12 |
141 | __movmemSI12: |
142 | mov.l @(8,r5),r0 |
143 | mov.l r0,@(8,r4) |
144 | .global __movmemSI8 |
145 | .global __movstrSI8 |
146 | .set __movstrSI8, __movmemSI8 |
147 | __movmemSI8: |
148 | mov.l @(4,r5),r0 |
149 | mov.l r0,@(4,r4) |
150 | .global __movmemSI4 |
151 | .global __movstrSI4 |
152 | .set __movstrSI4, __movmemSI4 |
153 | __movmemSI4: |
154 | mov.l @(0,r5),r0 |
155 | rts |
156 | mov.l r0,@(0,r4) |
157 | |
158 | .global __movmem_i4_even |
159 | .global __movstr_i4_even |
160 | .set __movstr_i4_even, __movmem_i4_even |
161 | |
162 | .global __movmem_i4_odd |
163 | .global __movstr_i4_odd |
164 | .set __movstr_i4_odd, __movmem_i4_odd |
165 | |
166 | .global __movmemSI12_i4 |
167 | .global __movstrSI12_i4 |
168 | .set __movstrSI12_i4, __movmemSI12_i4 |
169 | |
170 | .p2align 5 |
171 | L_movmem_2mod4_end: |
172 | mov.l r0,@(16,r4) |
173 | rts |
174 | mov.l r1,@(20,r4) |
175 | |
176 | .p2align 2 |
177 | |
178 | __movmem_i4_even: |
179 | mov.l @r5+,r0 |
180 | bra L_movmem_start_even |
181 | mov.l @r5+,r1 |
182 | |
183 | __movmem_i4_odd: |
184 | mov.l @r5+,r1 |
185 | add #-4,r4 |
186 | mov.l @r5+,r2 |
187 | mov.l @r5+,r3 |
188 | mov.l r1,@(4,r4) |
189 | mov.l r2,@(8,r4) |
190 | |
191 | L_movmem_loop: |
192 | mov.l r3,@(12,r4) |
193 | dt r6 |
194 | mov.l @r5+,r0 |
195 | bt/s L_movmem_2mod4_end |
196 | mov.l @r5+,r1 |
197 | add #16,r4 |
198 | L_movmem_start_even: |
199 | mov.l @r5+,r2 |
200 | mov.l @r5+,r3 |
201 | mov.l r0,@r4 |
202 | dt r6 |
203 | mov.l r1,@(4,r4) |
204 | bf/s L_movmem_loop |
205 | mov.l r2,@(8,r4) |
206 | rts |
207 | mov.l r3,@(12,r4) |
208 | |
209 | .p2align 4 |
210 | __movmemSI12_i4: |
211 | mov.l @r5,r0 |
212 | mov.l @(4,r5),r1 |
213 | mov.l @(8,r5),r2 |
214 | mov.l r0,@r4 |
215 | mov.l r1,@(4,r4) |
216 | rts |
217 | mov.l r2,@(8,r4) |
218 | |