rockbox/firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S
Jens Arnold d1e2f5240c c200: First implemention of lcd_yuv_blit(). mpegplayer works now, but speed can be optimized more and dithering is not yet implemented.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15020 a1c6a512-1295-4272-9138-f99709370657
2007-10-07 17:54:49 +00:00

243 lines
12 KiB
ArmAsm

/***************************************************************************
* __________ __ ___.
* Open \______ \ ____ ____ | | _\_ |__ _______ ___
* Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
* Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
* Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
* \/ \/ \/ \/ \/
* $Id$
*
* Copyright (C) 2007 by Jens Arnold
* Heavily based on lcd-as-memframe.c 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(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
*
* Write four RGB565 pixels in the following order on each loop:
* 1 3 + > down
* 2 4 \/ left
*/
.section .icode, "ax", %progbits
.align 2
.global lcd_write_yuv420_lines
.type lcd_write_yuv420_lines, %function
lcd_write_yuv420_lines:
@ r0 = yuv_src
@ r1 = width
@ r2 = stride
stmfd sp!, { r4-r12 } @ save non-scratch
ldmia r0, { r4, r5, r6 } @ r4 = yuv_src[0] = Y'_p
@ r5 = yuv_src[1] = Cb_p
@ r6 = yuv_src[2] = Cr_p
@ r0 = scratch
sub r2, r2, #1 @
mov r3, #0x70000000 @
orr r3, r3, #0x3000 @ r3 = LCD1_BASE
10: @ loop line @
ldrb r7, [r4], #1 @ r7 = *Y'_p++;
ldrb r8, [r5], #1 @ r8 = *Cb_p++;
ldrb r9, [r6], #1 @ r9 = *Cr_p++;
@
sub r7, r7, #16 @ r7 = Y = (Y' - 16)*74
add r12, r7, r7, asl #2 @ actually (Y' - 16)*37 and shift right
add r7, r12, r7, asl #5 @ by one less when adding - same for all
@
sub r8, r8, #128 @ Cb -= 128
sub r9, r9, #128 @ Cr -= 128
@
add r10, r9, r9, asl #1 @ r10 = Cr*51 + Cb*24
add r10, r10, r10, asl #4 @
add r10, r10, r8, asl #3 @
add r10, r10, r8, asl #4 @
@
add r11, r9, r9, asl #2 @ r9 = Cr*101
add r11, r11, r9, asl #5 @
add r9, r11, r9, asl #6 @
@
add r8, r8, #2 @ r8 = bu = (Cb*128 + 128) >> 8
mov r8, r8, asr #2 @
add r9, r9, #256 @ r9 = rv = (r9 + 256) >> 9
mov r9, r9, asr #9 @
rsb r10, r10, #128 @ r10 = guv = (-r10 + 128) >> 8
mov r10, r10, asr #8 @
@ compute R, G, and B
add r0, r8, r7, asr #8 @ r0 = b = (Y >> 9) + bu
add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv
@
orr r12, r0, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 15f @ no clamp @
cmp r0, #31 @ clamp b
mvnhi r0, r0, asr #31 @
andhi r0, r0, #31 @
cmp r11, #31 @ clamp r
mvnhi r11, r11, asr #31 @
andhi r11, r11, #31 @
cmp r7, #63 @ clamp g
mvnhi r7, r7, asr #31 @
andhi r7, r7, #63 @
15: @ no clamp @
@
ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride)
@
mov r11, r11, lsl #3 @
orr r11, r11, r7, lsr #3 @ r11 = (r << 3) | (g >> 3)
orr r0, r0, r7, lsl #5 @ r0 = (g << 5) | b
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r11, [r3, #0x10] @ send MSB
1: @busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r0, [r3, #0x10] @ send LSB
@
sub r7, r12, #16 @ r7 = Y = (Y' - 16)*74
add r12, r7, r7, asl #2 @
add r7, r12, r7, asl #5 @
@ compute R, G, and B
add r0, r8, r7, asr #8 @ r0 = b = (Y >> 9) + bu
add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv
@
orr r12, r0, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 15f @ no clamp @
cmp r0, #31 @ clamp b
mvnhi r0, r0, asr #31 @
andhi r0, r0, #31 @
cmp r11, #31 @ clamp r
mvnhi r11, r11, asr #31 @
andhi r11, r11, #31 @
cmp r7, #63 @ clamp g
mvnhi r7, r7, asr #31 @
andhi r7, r7, #63 @
15: @ no clamp @
@
ldrb r12, [r4], #1 @ r12 = Y' = *(Y'_p++)
@
mov r11, r11, lsl #3 @
orr r11, r11, r7, lsr #3 @ r11 = (r << 3) | (g >> 3)
orr r0, r0, r7, lsl #5 @ r0 = (g << 5) | b
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r11, [r3, #0x10] @ send MSB
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r0, [r3, #0x10] @ send LSB
@
sub r7, r12, #16 @ r7 = Y = (Y' - 16)*74
add r12, r7, r7, asl #2 @
add r7, r12, r7, asl #5 @
@ compute R, G, and B
add r0, r8, r7, asr #8 @ r0 = b = (Y >> 9) + bu
add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv
@
orr r12, r0, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 15f @ no clamp @
cmp r0, #31 @ clamp b
mvnhi r0, r0, asr #31 @
andhi r0, r0, #31 @
cmp r11, #31 @ clamp r
mvnhi r11, r11, asr #31 @
andhi r11, r11, #31 @
cmp r7, #63 @ clamp g
mvnhi r7, r7, asr #31 @
andhi r7, r7, #63 @
15: @ no clamp @
@
ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride)
@
@
mov r11, r11, lsl #3 @
orr r11, r11, r7, lsr #3 @ r11 = (r << 3) | (g >> 3)
orr r0, r0, r7, lsl #5 @ r0 = (g << 5) | b
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r11, [r3, #0x10] @ send MSB
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r0, [r3, #0x10] @ send LSB
@
sub r7, r12, #16 @ r7 = Y = (Y' - 16)*74
add r12, r7, r7, asl #2 @
add r7, r12, r7, asl #5 @
@ compute R, G, and B
add r0, r8, r7, asr #8 @ r0 = b = (Y >> 9) + bu
add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv
add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv
@
orr r12, r0, r11 @ check if clamping is needed...
orr r12, r12, r7, asr #1 @ ...at all
cmp r12, #31 @
bls 15f @ no clamp @
cmp r0, #31 @ clamp b
mvnhi r0, r0, asr #31 @
andhi r0, r0, #31 @
cmp r11, #31 @ clamp r
mvnhi r11, r11, asr #31 @
andhi r11, r11, #31 @
cmp r7, #63 @ clamp g
mvnhi r7, r7, asr #31 @
andhi r7, r7, #63 @
15: @ no clamp @
@
mov r11, r11, lsl #3 @
orr r11, r11, r7, lsr #3 @ r11 = (r << 3) | (g >> 3)
orr r0, r0, r7, lsl #5 @ r0 = (g << 5) | b
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r11, [r3, #0x10] @ send MSB
1: @ busy @
ldr r7, [r3] @ r7 = LCD1_BASE
tst r7, #LCD1_BUSY_MASK @ bridge busy?
bne 1b @
str r0, [r3, #0x10] @ send LSB
@
subs r1, r1, #2 @ subtract block from width
bgt 10b @ loop line @
@
ldmfd sp!, { r4-r12 } @ restore registers and return
bx lr @
.ltorg @ dump constant pool
.size lcd_write_yuv420_lines, .-lcd_write_yuv420_lines