diff options
author | Andree Buschmann <AndreeBuschmann@t-online.de> | 2010-12-29 23:17:47 +0000 |
---|---|---|
committer | Andree Buschmann <AndreeBuschmann@t-online.de> | 2010-12-29 23:17:47 +0000 |
commit | b04d676706c6e306754a3d1223d52de0037638bf (patch) | |
tree | bf2e34a68ca26c6c2397c51c91d4e06801d009e3 /firmware/target/arm/ipod/lcd-as-color-nano.S | |
parent | 1980fc3a61db6b85760ff044900094a231568936 (diff) | |
download | rockbox-b04d676706c6e306754a3d1223d52de0037638bf.tar.gz rockbox-b04d676706c6e306754a3d1223d52de0037638bf.zip |
Speed up of iPod nano 1G and iPod color LCD. Use HDD6330 asm part for YUV blitting, introduce special handling for full width screen updates. Speed up is about +30% for YUV on both color/nano1G.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@28930 a1c6a512-1295-4272-9138-f99709370657
Diffstat (limited to 'firmware/target/arm/ipod/lcd-as-color-nano.S')
-rwxr-xr-x | firmware/target/arm/ipod/lcd-as-color-nano.S | 152 |
1 files changed, 152 insertions, 0 deletions
diff --git a/firmware/target/arm/ipod/lcd-as-color-nano.S b/firmware/target/arm/ipod/lcd-as-color-nano.S new file mode 100755 index 0000000000..d4df4d496a --- /dev/null +++ b/firmware/target/arm/ipod/lcd-as-color-nano.S | |||
@@ -0,0 +1,152 @@ | |||
1 | /*************************************************************************** | ||
2 | * __________ __ ___. | ||
3 | * Open \______ \ ____ ____ | | _\_ |__ _______ ___ | ||
4 | * Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ / | ||
5 | * Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < < | ||
6 | * Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \ | ||
7 | * \/ \/ \/ \/ \/ | ||
8 | * $Id:$ | ||
9 | * | ||
10 | * Copyright (C) 2010 by Andree Buschmann | ||
11 | * | ||
12 | * Generic asm helper function used by YUV blitting. | ||
13 | * | ||
14 | * This program is free software; you can redistribute it and/or | ||
15 | * modify it under the terms of the GNU General Public License | ||
16 | * as published by the Free Software Foundation; either version 2 | ||
17 | * of the License, or (at your option) any later version. | ||
18 | * | ||
19 | * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY | ||
20 | * KIND, either express or implied. | ||
21 | * | ||
22 | ****************************************************************************/ | ||
23 | |||
24 | #include "config.h" | ||
25 | #include "cpu.h" | ||
26 | |||
27 | .section .icode, "ax", %progbits | ||
28 | |||
29 | /**************************************************************************** | ||
30 | * void lcd_yuv_write_inner_loop(unsigned char const * const ysrc, | ||
31 | * unsigned char const * const usrc, | ||
32 | * unsigned char const * const vsrc, | ||
33 | * int width); | ||
34 | * | ||
35 | * YUV- > RGB565 conversion | ||
36 | * |R| |1.000000 -0.000001 1.402000| |Y'| | ||
37 | * |G| = |1.000000 -0.334136 -0.714136| |Pb| | ||
38 | * |B| |1.000000 1.772000 0.000000| |Pr| | ||
39 | * Scaled, normalized, rounded and tweaked to yield RGB 565: | ||
40 | * |R| |74 0 101| |Y' - 16| >> 9 | ||
41 | * |G| = |74 -24 -51| |Cb - 128| >> 8 | ||
42 | * |B| |74 128 0| |Cr - 128| >> 9 | ||
43 | * | ||
44 | */ | ||
45 | .align 2 | ||
46 | .global lcd_yuv_write_inner_loop | ||
47 | .type lcd_yuv_write_inner_loop, %function | ||
48 | |||
49 | lcd_yuv_write_inner_loop: | ||
50 | @ r0 = ysrc | ||
51 | @ r1 = usrc | ||
52 | @ r2 = vsrc | ||
53 | @ r3 = width | ||
54 | stmfd sp!, { r4-r11, lr } @ save regs | ||
55 | mov r4, #0x70000000 @ r4 = LCD2_BLOCK_CTRL - 0x20 | ||
56 | add r4, r4, #0x8a00 @ | ||
57 | add r5, r4, #0x100 @ r5 = LCD2_BLOCK_DATA | ||
58 | 10: @ loop | ||
59 | |||
60 | ldrb r7, [r1], #1 @ *usrc++ | ||
61 | ldrb r8, [r2], #1 @ *vsrc++ | ||
62 | |||
63 | sub r7, r7, #128 @ Cb -= 128 | ||
64 | sub r8, r8, #128 @ Cr -= 128 | ||
65 | |||
66 | add r10, r8, r8, asl #2 @ Cr*101 | ||
67 | add r10, r10, r8, asl #5 | ||
68 | add r10, r10, r8, asl #6 | ||
69 | |||
70 | add r11, r8, r8, asl #1 @ Cr*51 + Cb*24 | ||
71 | add r11, r11, r11, asl #4 | ||
72 | add r11, r11, r7, asl #3 | ||
73 | add r11, r11, r7, asl #4 | ||
74 | |||
75 | add r12, r7, #2 @ r12 = bu = (Cb*128 + 256) >> 9 | ||
76 | mov r12, r12, asr #2 | ||
77 | add r10, r10, #256 @ r10 = rv = (Cr*101 + 256) >> 9 | ||
78 | mov r10, r10, asr #9 | ||
79 | rsb r11, r11, #128 @ r11 = guv = (-r11 + 128) >> 8 | ||
80 | mov r11, r11, asr #8 | ||
81 | |||
82 | @ pixel_1 | ||
83 | ldrb r7, [r0], #1 @ *ysrc++ | ||
84 | sub r7, r7, #16 @ Y = (Y' - 16) * 37 | ||
85 | add r8, r7, r7, asl #2 | ||
86 | add r7, r8, r7, asl #5 | ||
87 | |||
88 | add r9, r10, r7, asr #8 @ R = (Y >> 8) + rv | ||
89 | add r8, r11, r7, asr #7 @ G = (Y >> 7) + guv | ||
90 | add r7, r12, r7, asr #8 @ B = (Y >> 8) + bu | ||
91 | |||
92 | cmp r9, #31 @ clamp R | ||
93 | mvnhi r9, r9, asr #31 | ||
94 | andhi r9, r9, #31 | ||
95 | |||
96 | cmp r8, #63 @ clamp G | ||
97 | mvnhi r8, r8, asr #31 | ||
98 | andhi r8, r8, #63 | ||
99 | |||
100 | cmp r7, #31 @ clamp B | ||
101 | mvnhi r7, r7, asr #31 | ||
102 | andhi r7, r7, #31 | ||
103 | |||
104 | orr r6, r7, r8, lsl #5 @ pack pixel | ||
105 | orr r6, r6, r9, lsl #11 | ||
106 | |||
107 | mov r7, r6, lsl #8 @ swap bytes | ||
108 | and r7, r7, #0xff00 | ||
109 | add r6, r7, r6, lsr #8 | ||
110 | |||
111 | @ pixel_2 | ||
112 | ldrb r7, [r0], #1 @ *ysrc++ | ||
113 | sub r7, r7, #16 @ Y = (Y' - 16) * 37 | ||
114 | add r8, r7, r7, asl #2 | ||
115 | add r7, r8, r7, asl #5 | ||
116 | |||
117 | add r9, r10, r7, asr #8 @ R = (Y >> 8) + rv | ||
118 | add r8, r11, r7, asr #7 @ G = (Y >> 7) + guv | ||
119 | add r7, r12, r7, asr #8 @ B = (Y >> 8) + bu | ||
120 | |||
121 | cmp r9, #31 @ clamp R | ||
122 | mvnhi r9, r9, asr #31 | ||
123 | andhi r9, r9, #31 | ||
124 | |||
125 | cmp r8, #63 @ clamp G | ||
126 | mvnhi r8, r8, asr #31 | ||
127 | andhi r8, r8, #63 | ||
128 | |||
129 | cmp r7, #31 @ clamp B | ||
130 | mvnhi r7, r7, asr #31 | ||
131 | andhi r7, r7, #31 | ||
132 | |||
133 | orr r7, r7, r8, lsl #5 @ pack pixel | ||
134 | orr r7, r7, r9, lsl #11 | ||
135 | |||
136 | orr r6, r6, r7, lsl #24 @ swap bytes and add pixels simultaneously | ||
137 | mov r7, r7, lsr #8 | ||
138 | orr r6, r6, r7, lsl #16 | ||
139 | #if 1 | ||
140 | 11: @ while (!(LCD2_BLOCK_CTRL & LCD2_BLOCK_TXOK)); | ||
141 | ldr r11, [r4, #0x20] @ | ||
142 | tst r11, #0x1000000 @ | ||
143 | beq 11b @ | ||
144 | #endif | ||
145 | str r6, [r5] @ send two pixels | ||
146 | |||
147 | subs r3, r3, #2 @ decrease width | ||
148 | bgt 10b @ loop | ||
149 | |||
150 | ldmpc regs=r4-r11 @ restore regs | ||
151 | .ltorg @ dump constant pool | ||
152 | .size lcd_yuv_write_inner_loop, .-lcd_yuv_write_inner_loop | ||