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
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
|
/***************************************************************************
* __________ __ ___.
* Open \______ \ ____ ____ | | _\_ |__ _______ ___
* Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
* Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
* Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
* \/ \/ \/ \/ \/
* $Id$
*
* Copyright (C) 2007 by Michael Sevakis
*
* All files in this archive are subject to the GNU General Public License.
* See the file COPYING in the source tree root for full license agreement.
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
****************************************************************************/
#include "config.h"
#include "cpu.h"
/****************************************************************************
* void lcd_write_yuv_420_lines(fb_data *dst,
* unsigned char chroma_buf[LCD_HEIGHT/2*3],
unsigned char const * const src[3],
* int width,
* int stride);
*
* |R| |1.000000 -0.000001 1.402000| |Y'|
* |G| = |1.000000 -0.334136 -0.714136| |Pb|
* |B| |1.000000 1.772000 0.000000| |Pr|
* Scaled, normalized, rounded and tweaked to yield RGB 565:
* |R| |74 0 101| |Y' - 16| >> 9
* |G| = |74 -24 -51| |Cb - 128| >> 8
* |B| |74 128 0| |Cr - 128| >> 9
*/
.section .icode, "ax", %progbits
.align 2
.global lcd_write_yuv420_lines
.type lcd_write_yuv420_lines, %function
lcd_write_yuv420_lines:
@ r0 = dst
@ r1 = chroma_buf
@ r2 = yuv_src
@ r3 = width
@ [sp] = stride
stmdb sp!, { r4-r12, lr } @ save non-scratch
stmdb sp!, { r0, r3 } @ save dst and width
mov r14, #74 @ r14 = Y factor
ldmia r2, { r4, r5, r6 } @ r4 = yuv_src[0] = Y'_p
@ r5 = yuv_src[1] = Cb_p
@ r6 = yuv_src[2] = Cr_p
10: @ loop line 1 @
ldrb r2, [r4], #1 @ r2 = *Y'_p++;
ldrb r8, [r5], #1 @ r8 = *Cb_p++;
ldrb r11, [r6], #1 @ r11 = *Cr_p++;
@
@ compute Y
sub r2, r2, #16 @ r7 = Y = (Y' - 16)*74
mul r7, r2, r14 @
@
sub r8, r8, #128 @ Cb -= 128
sub r11, r11, #128 @ Cr -= 128
@
mvn r2, #24 @ compute guv
mul r10, r2, r8 @ r10 = Cb*-24
mvn r2, #51 @
mla r10, r2, r11, r10 @ r10 = r10 + Cr*-51
@
mov r2, #101 @ compute rv
mul r9, r11, r2 @ r9 = rv = Cr*101
@
@ store chromas in line buffer
add r8, r8, #2 @ bu = (Cb + 2) >> 2
mov r8, r8, asr #2 @
strb r8, [r1], #1 @
add r9, r9, #256 @ rv = (Cr + 256) >> 9
mov r9, r9, asr #9 @
strb r9, [r1], #1 @
mov r10, r10, asr #8 @ guv >>= 8
strb r10, [r1], #1 @
@ compute R, G, and B
add r2, r8, r7, asr #9 @ r2 = b = (Y >> 9) + bu
add r11, r9, r7, asr #9 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #8 @ r7 = g = (Y >> 8) + guv
@
orr r12, r2, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 15f @ no clamp @
mov r12, #31 @
cmp r12, r2 @ clamp b
andlo r2, r12, r2, asr #31 @
eorlo r2, r2, r12 @
cmp r12, r11 @ clamp r
andlo r11, r12, r11, asr #31 @
eorlo r11, r11, r12 @
cmp r12, r7, asr #1 @ clamp g
andlo r7, r12, r7, asr #31 @
eorlo r7, r7, r12 @
orrlo r7, r7, r7, asl #1 @
15: @ no clamp @
@
orr r12, r2, r7, lsl #5 @ r4 |= (g << 5)
ldrb r2, [r4], #1 @ r2 = Y' = *Y'_p++
orr r12, r12, r11, lsl #11 @ r4 = b | (r << 11)
strh r12, [r0], #240 @ store pixel
@
sub r2, r2, #16 @ r7 = Y = (Y' - 16)*74
mul r7, r2, r14 @ next Y
@ compute R, G, and B
add r2, r8, r7, asr #9 @ r2 = b = (Y >> 9) + bu
add r11, r9, r7, asr #9 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #8 @ r7 = g = (Y >> 8) + guv
@
orr r12, r2, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 15f @ no clamp @
mov r12, #31 @
cmp r12, r2 @ clamp b
andlo r2, r12, r2, asr #31 @
eorlo r2, r2, r12 @
cmp r12, r11 @ clamp r
andlo r11, r12, r11, asr #31 @
eorlo r11, r11, r12 @
cmp r12, r7, asr #1 @ clamp g
andlo r7, r12, r7, asr #31 @
eorlo r7, r7, r12 @
orrlo r7, r7, r7, asl #1 @
15: @ no clamp @
@
orr r12, r2, r11, lsl #11 @ r4 = b | (r << 11)
orr r12, r12, r7, lsl #5 @ r4 |= (g << 5)
strh r12, [r0, #240]! @ store pixel
add r0, r0, #2*240 @
@
subs r3, r3, #2 @
bgt 10b @ loop line 1 @
@ do second line
@
ldmia sp!, { r0, r3 } @ pop dst and width
sub r0, r0, #2 @ set dst to start of next line
sub r1, r1, r3, asl #1 @ rewind chroma pointer...
ldr r2, [sp, #40] @ r2 = stride
add r1, r1, r3, asr #1 @ ... (r1 -= width/2*3)
@ move sources to start of next line
sub r2, r2, r3 @ r2 = skip = stride - width
add r4, r4, r2 @ r4 = Y'_p + skip
@
20: @ loop line 2 @
ldrb r2, [r4], #1 @ r7 = Y' = *Y'_p++
ldrsb r8, [r1], #1 @ reload saved chromas
ldrsb r9, [r1], #1 @
ldrsb r10, [r1], #1 @
@
sub r2, r2, #16 @ r2 = Y = (Y' - 16)*74
mul r7, r2, r14 @
@ compute R, G, and B
add r2, r8, r7, asr #9 @ r2 = b = (Y >> 9) + bu
add r11, r9, r7, asr #9 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #8 @ r7 = g = (Y >> 8) + guv
@
orr r12, r2, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 25f @ no clamp @
mov r12, #31 @
cmp r12, r2 @ clamp b
andlo r2, r12, r2, asr #31 @
eorlo r2, r2, r12 @
cmp r12, r11 @ clamp r
andlo r11, r12, r11, asr #31 @
eorlo r11, r11, r12 @
cmp r12, r7, asr #1 @ clamp g
andlo r7, r12, r7, asr #31 @
eorlo r7, r7, r12 @
orrlo r7, r7, r7, asl #1 @
25: @ no clamp @
@
orr r12, r2, r11, lsl #11 @ r4 = b | (r << 11)
ldrb r2, [r4], #1 @ r2 = Y' = *Y'_p++
orr r12, r12, r7, lsl #5 @ r4 |= (g << 5)
strh r12, [r0], #240 @ store pixel
@
@ do second pixel
@
sub r2, r2, #16 @ r2 = Y = (Y' - 16)*74
mul r7, r2, r14 @
@ compute R, G, and B
add r2, r8, r7, asr #9 @ r2 = b = (Y >> 9) + bu
add r11, r9, r7, asr #9 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #8 @ r7 = g = (Y >> 8) + guv
@
orr r12, r2, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 25f @ no clamp @
mov r12, #31 @
cmp r12, r2 @ clamp b
andlo r2, r12, r2, asr #31 @
eorlo r2, r2, r12 @
cmp r12, r11 @ clamp r
andlo r11, r12, r11, asr #31 @
eorlo r11, r11, r12 @
cmp r12, r7, asr #1 @ clamp g
andlo r7, r12, r7, asr #31 @
eorlo r7, r7, r12 @
orrlo r7, r7, r7, asl #1 @
25: @ no clamp @
@
orr r12, r2, r11, lsl #11 @ r4 = b | (r << 11)
orr r12, r12, r7, lsl #5 @ r4 |= (g << 5)
strh r12, [r0, #240]! @ store pixel
add r0, r0, #2*240 @
@
subs r3, r3, #2 @
bgt 20b @ loop line 2 @
@
ldmia sp!, { r4-r12, pc } @ restore registers and return
.size lcd_write_yuv420_lines, .-lcd_write_yuv420_lines
|
