diff options
Diffstat (limited to 'firmware/target/arm/pbell/vibe500/lcd-as-vibe500.S')
| -rw-r--r-- | firmware/target/arm/pbell/vibe500/lcd-as-vibe500.S | 556 |
1 files changed, 556 insertions, 0 deletions
diff --git a/firmware/target/arm/pbell/vibe500/lcd-as-vibe500.S b/firmware/target/arm/pbell/vibe500/lcd-as-vibe500.S new file mode 100644 index 0000000000..e03011c168 --- /dev/null +++ b/firmware/target/arm/pbell/vibe500/lcd-as-vibe500.S | |||
| @@ -0,0 +1,556 @@ | |||
| 1 | /*************************************************************************** | ||
| 2 | * __________ __ ___. | ||
| 3 | * Open \______ \ ____ ____ | | _\_ |__ _______ ___ | ||
| 4 | * Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ / | ||
| 5 | * Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < < | ||
| 6 | * Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \ | ||
| 7 | * \/ \/ \/ \/ \/ | ||
| 8 | * $Id:$ | ||
| 9 | * | ||
| 10 | * Copyright (C) 2007-2008 by Michael Sevakis | ||
| 11 | * Adapted for the Packard Bell Vibe 500 by Szymon Dziok | ||
| 12 | * | ||
| 13 | * Packard Bell Vibe 500 LCD assembly routines | ||
| 14 | * | ||
| 15 | * This program is free software; you can redistribute it and/or | ||
| 16 | * modify it under the terms of the GNU General Public License | ||
| 17 | * as published by the Free Software Foundation; either version 2 | ||
| 18 | * of the License, or (at your option) any later version. | ||
| 19 | * | ||
| 20 | * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY | ||
| 21 | * KIND, either express or implied. | ||
| 22 | * | ||
| 23 | ****************************************************************************/ | ||
| 24 | |||
| 25 | #include "config.h" | ||
| 26 | #include "cpu.h" | ||
| 27 | |||
| 28 | /**************************************************************************** | ||
| 29 | * void lcd_write_yuv420_lines(unsigned char const * const src[3], | ||
| 30 | * int width, | ||
| 31 | * int stride); | ||
| 32 | * | ||
| 33 | * |R| |1.000000 -0.000001 1.402000| |Y'| | ||
| 34 | * |G| = |1.000000 -0.334136 -0.714136| |Pb| | ||
| 35 | * |B| |1.000000 1.772000 0.000000| |Pr| | ||
| 36 | * Scaled, normalized, rounded and tweaked to yield RGB 565: | ||
| 37 | * |R| |74 0 101| |Y' - 16| >> 9 | ||
| 38 | * |G| = |74 -24 -51| |Cb - 128| >> 8 | ||
| 39 | * |B| |74 128 0| |Cr - 128| >> 9 | ||
| 40 | * | ||
| 41 | * Write four RGB565 pixels in the following order on each loop: | ||
| 42 | * 1 3 + > down | ||
| 43 | * 2 4 \/ left | ||
| 44 | */ | ||
| 45 | .section .icode, "ax", %progbits | ||
| 46 | .align 2 | ||
| 47 | .global lcd_write_yuv420_lines | ||
| 48 | .type lcd_write_yuv420_lines, %function | ||
| 49 | lcd_write_yuv420_lines: | ||
| 50 | @ r0 = yuv_src | ||
| 51 | @ r1 = width | ||
| 52 | @ r2 = stride | ||
| 53 | stmfd sp!, { r4-r11, lr } @ save non-scratch | ||
| 54 | ldmia r0, { r4, r5, r6 } @ r4 = yuv_src[0] = Y'_p | ||
| 55 | @ r5 = yuv_src[1] = Cb_p | ||
| 56 | @ r6 = yuv_src[2] = Cr_p | ||
| 57 | @ | ||
| 58 | ldr r0, =LCD1_BASE @ | ||
| 59 | @ | ||
| 60 | sub r2, r2, #1 @ Adjust stride because of increment | ||
| 61 | 10: @ loop line @ | ||
| 62 | ldrb r7, [r4], #1 @ r7 = *Y'_p++; | ||
| 63 | ldrb r8, [r5], #1 @ r8 = *Cb_p++; | ||
| 64 | ldrb r9, [r6], #1 @ r9 = *Cr_p++; | ||
| 65 | @ | ||
| 66 | sub r7, r7, #16 @ r7 = Y = (Y' - 16)*74 | ||
| 67 | add r12, r7, r7, asl #2 @ actually (Y' - 16)*37 and shift right | ||
| 68 | add r7, r12, r7, asl #5 @ by one less when adding - same for all | ||
| 69 | @ | ||
| 70 | sub r8, r8, #128 @ Cb -= 128 | ||
| 71 | sub r9, r9, #128 @ Cr -= 128 | ||
| 72 | @ | ||
| 73 | add r10, r9, r9, asl #1 @ r10 = Cr*51 + Cb*24 | ||
| 74 | add r10, r10, r10, asl #4 @ | ||
| 75 | add r10, r10, r8, asl #3 @ | ||
| 76 | add r10, r10, r8, asl #4 @ | ||
| 77 | @ | ||
| 78 | add r11, r9, r9, asl #2 @ r9 = Cr*101 | ||
| 79 | add r11, r11, r9, asl #5 @ | ||
| 80 | add r9, r11, r9, asl #6 @ | ||
| 81 | @ | ||
| 82 | add r8, r8, #2 @ r8 = bu = (Cb*128 + 128) >> 8 | ||
| 83 | mov r8, r8, asr #2 @ | ||
| 84 | add r9, r9, #256 @ r9 = rv = (r8 + 256) >> 9 | ||
| 85 | mov r9, r9, asr #9 @ | ||
| 86 | rsb r10, r10, #128 @ r10 = guv = (-r9 + 128) >> 8 | ||
| 87 | mov r10, r10, asr #8 @ | ||
| 88 | @ compute R, G, and B | ||
| 89 | add r3, r8, r7, asr #8 @ r3 = b = (Y >> 9) + bu | ||
| 90 | add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv | ||
| 91 | add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv | ||
| 92 | @ | ||
| 93 | orr r12, r3, r11 @ check if clamping is needed... | ||
| 94 | orr r12, r12, r7, asr #1 @ ...at all | ||
| 95 | cmp r12, #31 @ | ||
| 96 | bls 15f @ no clamp @ | ||
| 97 | cmp r3, #31 @ clamp b | ||
| 98 | mvnhi r3, r3, asr #31 @ | ||
| 99 | andhi r3, r3, #31 @ | ||
| 100 | cmp r11, #31 @ clamp r | ||
| 101 | mvnhi r11, r11, asr #31 @ | ||
| 102 | andhi r11, r11, #31 @ | ||
| 103 | cmp r7, #63 @ clamp g | ||
| 104 | mvnhi r7, r7, asr #31 @ | ||
| 105 | andhi r7, r7, #63 @ | ||
| 106 | 15: @ no clamp @ | ||
| 107 | @ | ||
| 108 | ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride) | ||
| 109 | @ | ||
| 110 | orr r3, r3, r11, lsl #11 @ r3 = b | (r << 11) | ||
| 111 | orr r3, r3, r7, lsl #5 @ r3 |= (g << 5) | ||
| 112 | @ | ||
| 113 | movs r7, r3, lsr #8 @ store pixel | ||
| 114 | 20: @ | ||
| 115 | ldr r11, [r0] @ | ||
| 116 | tst r11, #LCD1_BUSY_MASK @ | ||
| 117 | bne 20b @ | ||
| 118 | str r7, [r0, #0x10] @ | ||
| 119 | 25: @ | ||
| 120 | ldr r11, [r0] @ | ||
| 121 | tst r11, #LCD1_BUSY_MASK @ | ||
| 122 | bne 25b @ | ||
| 123 | str r3, [r0, #0x10] @ | ||
| 124 | @ | ||
| 125 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*74 | ||
| 126 | add r12, r7, r7, asl #2 @ | ||
| 127 | add r7, r12, r7, asl #5 @ | ||
| 128 | @ compute R, G, and B | ||
| 129 | add r3, r8, r7, asr #8 @ r3 = b = (Y >> 9) + bu | ||
| 130 | add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv | ||
| 131 | add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv | ||
| 132 | @ | ||
| 133 | orr r12, r3, r11 @ check if clamping is needed... | ||
| 134 | orr r12, r12, r7, asr #1 @ ...at all | ||
| 135 | cmp r12, #31 @ | ||
| 136 | bls 15f @ no clamp @ | ||
| 137 | cmp r3, #31 @ clamp b | ||
| 138 | mvnhi r3, r3, asr #31 @ | ||
| 139 | andhi r3, r3, #31 @ | ||
| 140 | cmp r11, #31 @ clamp r | ||
| 141 | mvnhi r11, r11, asr #31 @ | ||
| 142 | andhi r11, r11, #31 @ | ||
| 143 | cmp r7, #63 @ clamp g | ||
| 144 | mvnhi r7, r7, asr #31 @ | ||
| 145 | andhi r7, r7, #63 @ | ||
| 146 | 15: @ no clamp @ | ||
| 147 | @ | ||
| 148 | ldrb r12, [r4], #1 @ r12 = Y' = *(Y'_p++) | ||
| 149 | @ | ||
| 150 | orr r3, r3, r11, lsl #11 @ r3 = b | (r << 11) | ||
| 151 | orr r3, r3, r7, lsl #5 @ r3 |= (g << 5) | ||
| 152 | @ | ||
| 153 | movs r7, r3, lsr #8 @ store pixel | ||
| 154 | 20: @ | ||
| 155 | ldr r11, [r0] @ | ||
| 156 | tst r11, #LCD1_BUSY_MASK @ | ||
| 157 | bne 20b @ | ||
| 158 | str r7, [r0, #0x10] @ | ||
| 159 | 25: @ | ||
| 160 | ldr r11, [r0] @ | ||
| 161 | tst r11, #LCD1_BUSY_MASK @ | ||
| 162 | bne 25b @ | ||
| 163 | str r3, [r0, #0x10] @ | ||
| 164 | @ | ||
| 165 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*74 | ||
| 166 | add r12, r7, r7, asl #2 @ | ||
| 167 | add r7, r12, r7, asl #5 @ | ||
| 168 | @ compute R, G, and B | ||
| 169 | add r3, r8, r7, asr #8 @ r3 = b = (Y >> 9) + bu | ||
| 170 | add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv | ||
| 171 | add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv | ||
| 172 | @ | ||
| 173 | orr r12, r3, r11 @ check if clamping is needed... | ||
| 174 | orr r12, r12, r7, asr #1 @ ...at all | ||
| 175 | cmp r12, #31 @ | ||
| 176 | bls 15f @ no clamp @ | ||
| 177 | cmp r3, #31 @ clamp b | ||
| 178 | mvnhi r3, r3, asr #31 @ | ||
| 179 | andhi r3, r3, #31 @ | ||
| 180 | cmp r11, #31 @ clamp r | ||
| 181 | mvnhi r11, r11, asr #31 @ | ||
| 182 | andhi r11, r11, #31 @ | ||
| 183 | cmp r7, #63 @ clamp g | ||
| 184 | mvnhi r7, r7, asr #31 @ | ||
| 185 | andhi r7, r7, #63 @ | ||
| 186 | 15: @ no clamp @ | ||
| 187 | @ | ||
| 188 | ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride) | ||
| 189 | @ | ||
| 190 | orr r3, r3, r7, lsl #5 @ r3 = b | (g << 5) | ||
| 191 | orr r3, r3, r11, lsl #11 @ r3 |= (r << 11) | ||
| 192 | @ | ||
| 193 | movs r7, r3, lsr #8 @ store pixel | ||
| 194 | 20: @ | ||
| 195 | ldr r11, [r0] @ | ||
| 196 | tst r11, #LCD1_BUSY_MASK @ | ||
| 197 | bne 20b @ | ||
| 198 | str r7, [r0, #0x10] @ | ||
| 199 | 25: @ | ||
| 200 | ldr r11, [r0] @ | ||
| 201 | tst r11, #LCD1_BUSY_MASK @ | ||
| 202 | bne 25b @ | ||
| 203 | str r3, [r0, #0x10] @ | ||
| 204 | @ | ||
| 205 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*74 | ||
| 206 | add r12, r7, r7, asl #2 @ | ||
| 207 | add r7, r12, r7, asl #5 @ | ||
| 208 | @ compute R, G, and B | ||
| 209 | add r3, r8, r7, asr #8 @ r3 = b = (Y >> 9) + bu | ||
| 210 | add r11, r9, r7, asr #8 @ r11 = r = (Y >> 9) + rv | ||
| 211 | add r7, r10, r7, asr #7 @ r7 = g = (Y >> 8) + guv | ||
| 212 | @ | ||
| 213 | orr r12, r3, r11 @ check if clamping is needed... | ||
| 214 | orr r12, r12, r7, asr #1 @ ...at all | ||
| 215 | cmp r12, #31 @ | ||
| 216 | bls 15f @ no clamp @ | ||
| 217 | cmp r3, #31 @ clamp b | ||
| 218 | mvnhi r3, r3, asr #31 @ | ||
| 219 | andhi r3, r3, #31 @ | ||
| 220 | cmp r11, #31 @ clamp r | ||
| 221 | mvnhi r11, r11, asr #31 @ | ||
| 222 | andhi r11, r11, #31 @ | ||
| 223 | cmp r7, #63 @ clamp g | ||
| 224 | mvnhi r7, r7, asr #31 @ | ||
| 225 | andhi r7, r7, #63 @ | ||
| 226 | 15: @ no clamp @ | ||
| 227 | @ | ||
| 228 | orr r3, r3, r11, lsl #11 @ r3 = b | (r << 11) | ||
| 229 | orr r3, r3, r7, lsl #5 @ r3 |= (g << 5) | ||
| 230 | @ | ||
| 231 | movs r7, r3, lsr #8 @ store pixel | ||
| 232 | 20: @ | ||
| 233 | ldr r11, [r0] @ | ||
| 234 | tst r11, #LCD1_BUSY_MASK @ | ||
| 235 | bne 20b @ | ||
| 236 | str r7, [r0, #0x10] @ | ||
| 237 | 25: @ | ||
| 238 | ldr r11, [r0] @ | ||
| 239 | tst r11, #LCD1_BUSY_MASK @ | ||
| 240 | bne 25b @ | ||
| 241 | str r3, [r0, #0x10] @ | ||
| 242 | @ | ||
| 243 | subs r1, r1, #2 @ subtract block from width | ||
| 244 | bgt 10b @ loop line @ | ||
| 245 | @ | ||
| 246 | ldmpc regs=r4-r11 @ restore registers and return | ||
| 247 | .ltorg @ dump constant pool | ||
| 248 | .size lcd_write_yuv420_lines, .-lcd_write_yuv420_lines | ||
| 249 | |||
| 250 | |||
| 251 | /**************************************************************************** | ||
| 252 | * void lcd_write_yuv420_lines_odither(unsigned char const * const src[3], | ||
| 253 | * int width, | ||
| 254 | * int stride, | ||
| 255 | * int x_screen, | ||
| 256 | * int y_screen); | ||
| 257 | * | ||
| 258 | * |R| |1.000000 -0.000001 1.402000| |Y'| | ||
| 259 | * |G| = |1.000000 -0.334136 -0.714136| |Pb| | ||
| 260 | * |B| |1.000000 1.772000 0.000000| |Pr| | ||
| 261 | * Red scaled at twice g & b but at same precision to place it in correct | ||
| 262 | * bit position after multiply and leave instruction count lower. | ||
| 263 | * |R| |258 0 408| |Y' - 16| | ||
| 264 | * |G| = |149 -49 -104| |Cb - 128| | ||
| 265 | * |B| |149 258 0| |Cr - 128| | ||
| 266 | * | ||
| 267 | * Write four RGB565 pixels in the following order on each loop: | ||
| 268 | * 1 3 + > down | ||
| 269 | * 2 4 \/ left | ||
| 270 | * | ||
| 271 | * Kernel pattern (raw|use order): | ||
| 272 | * 5 3 4 2 row0 row2 > down | ||
| 273 | * 1 7 0 6 | 5 1 3 7 4 0 2 6 col0 left | ||
| 274 | * 4 2 5 3 | 4 0 2 6 5 1 3 7 col2 \/ | ||
| 275 | * 0 6 1 7 | ||
| 276 | */ | ||
| 277 | .section .icode, "ax", %progbits | ||
| 278 | .align 2 | ||
| 279 | .global lcd_write_yuv420_lines_odither | ||
| 280 | .type lcd_write_yuv420_lines_odither, %function | ||
| 281 | lcd_write_yuv420_lines_odither: | ||
| 282 | @ r0 = yuv_src | ||
| 283 | @ r1 = width | ||
| 284 | @ r2 = stride | ||
| 285 | @ r3 = x_screen | ||
| 286 | @ [sp] = y_screen | ||
| 287 | stmfd sp!, { r4-r11, lr } @ save non-scratch | ||
| 288 | ldmia r0, { r4, r5, r6 } @ r4 = yuv_src[0] = Y'_p | ||
| 289 | @ r5 = yuv_src[1] = Cb_p | ||
| 290 | @ r6 = yuv_src[2] = Cr_p | ||
| 291 | @ | ||
| 292 | ldr r0, [sp, #36] @ Line up pattern and kernel quadrant | ||
| 293 | eor r14, r3, r0 @ | ||
| 294 | and r14, r14, #0x2 @ | ||
| 295 | mov r14, r14, lsl #6 @ 0x00 or 0x80 | ||
| 296 | @ | ||
| 297 | ldr r0, =LCD1_BASE @ | ||
| 298 | @ | ||
| 299 | sub r2, r2, #1 @ Adjust stride because of increment | ||
| 300 | 10: @ loop line @ | ||
| 301 | @ | ||
| 302 | ldrb r7, [r4], #1 @ r7 = *Y'_p++; | ||
| 303 | ldrb r8, [r5], #1 @ r8 = *Cb_p++; | ||
| 304 | ldrb r9, [r6], #1 @ r9 = *Cr_p++; | ||
| 305 | @ | ||
| 306 | eor r14, r14, #0x80 @ flip pattern quadrant | ||
| 307 | @ | ||
| 308 | sub r7, r7, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 309 | add r12, r7, r7, asl #2 @ | ||
| 310 | add r12, r12, r12, asl #4 @ | ||
| 311 | add r7, r12, r7, asl #6 @ | ||
| 312 | @ | ||
| 313 | sub r8, r8, #128 @ Cb -= 128 | ||
| 314 | sub r9, r9, #128 @ Cr -= 128 | ||
| 315 | @ | ||
| 316 | add r10, r8, r8, asl #4 @ r10 = guv = Cr*104 + Cb*49 | ||
| 317 | add r10, r10, r8, asl #5 @ | ||
| 318 | add r10, r10, r9, asl #3 @ | ||
| 319 | add r10, r10, r9, asl #5 @ | ||
| 320 | add r10, r10, r9, asl #6 @ | ||
| 321 | @ | ||
| 322 | mov r8, r8, asl #1 @ r8 = bu = Cb*258 | ||
| 323 | add r8, r8, r8, asl #7 @ | ||
| 324 | @ | ||
| 325 | add r9, r9, r9, asl #1 @ r9 = rv = Cr*408 | ||
| 326 | add r9, r9, r9, asl #4 @ | ||
| 327 | mov r9, r9, asl #3 @ | ||
| 328 | @ | ||
| 329 | @ compute R, G, and B | ||
| 330 | add r3, r8, r7 @ r3 = b' = Y + bu | ||
| 331 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 332 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 333 | @ | ||
| 334 | @ r8 = bu, r9 = rv, r10 = guv | ||
| 335 | @ | ||
| 336 | sub r12, r3, r3, lsr #5 @ r3 = 31/32*b + b/256 | ||
| 337 | add r3, r12, r3, lsr #8 @ | ||
| 338 | @ | ||
| 339 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r + r/256 | ||
| 340 | add r11, r12, r11, lsr #8 @ | ||
| 341 | @ | ||
| 342 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g + g/256 | ||
| 343 | add r7, r12, r7, lsr #8 @ | ||
| 344 | @ | ||
| 345 | add r12, r14, #0x200 @ | ||
| 346 | @ | ||
| 347 | add r3, r3, r12 @ b = r3 + delta | ||
| 348 | add r11, r11, r12, lsl #1 @ r = r11 + delta*2 | ||
| 349 | add r7, r7, r12, lsr #1 @ g = r7 + delta/2 | ||
| 350 | @ | ||
| 351 | orr r12, r3, r11, asr #1 @ check if clamping is needed... | ||
| 352 | orr r12, r12, r7 @ ...at all | ||
| 353 | movs r12, r12, asr #15 @ | ||
| 354 | beq 15f @ no clamp @ | ||
| 355 | movs r12, r3, asr #15 @ clamp b | ||
| 356 | mvnne r3, r12, lsr #15 @ | ||
| 357 | andne r3, r3, #0x7c00 @ mask b only if clamped | ||
| 358 | movs r12, r11, asr #16 @ clamp r | ||
| 359 | mvnne r11, r12, lsr #16 @ | ||
| 360 | movs r12, r7, asr #15 @ clamp g | ||
| 361 | mvnne r7, r12, lsr #15 @ | ||
| 362 | 15: @ no clamp @ | ||
| 363 | @ | ||
| 364 | ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride) | ||
| 365 | @ | ||
| 366 | and r11, r11, #0xf800 @ pack pixel | ||
| 367 | and r7, r7, #0x7e00 @ r3 = pixel = (r & 0xf800) | | ||
| 368 | orr r11, r11, r7, lsr #4 @ ((g & 0x7e00) >> 4) | | ||
| 369 | orr r3, r11, r3, lsr #10 @ (b >> 10) | ||
| 370 | @ | ||
| 371 | movs r7, r3, lsr #8 @ store pixel | ||
| 372 | 20: @ | ||
| 373 | ldr r11, [r0] @ | ||
| 374 | tst r11, #LCD1_BUSY_MASK @ | ||
| 375 | bne 20b @ | ||
| 376 | str r7, [r0, #0x10] @ | ||
| 377 | 25: @ | ||
| 378 | ldr r11, [r0] @ | ||
| 379 | tst r11, #LCD1_BUSY_MASK @ | ||
| 380 | bne 25b @ | ||
| 381 | str r3, [r0, #0x10] @ | ||
| 382 | @ | ||
| 383 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 384 | add r12, r7, r7, asl #2 @ | ||
| 385 | add r12, r12, r12, asl #4 @ | ||
| 386 | add r7, r12, r7, asl #6 @ | ||
| 387 | @ compute R, G, and B | ||
| 388 | add r3, r8, r7 @ r3 = b' = Y + bu | ||
| 389 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 390 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 391 | @ | ||
| 392 | sub r12, r3, r3, lsr #5 @ r3 = 31/32*b' + b'/256 | ||
| 393 | add r3, r12, r3, lsr #8 @ | ||
| 394 | @ | ||
| 395 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r' + r'/256 | ||
| 396 | add r11, r12, r11, lsr #8 @ | ||
| 397 | @ | ||
| 398 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g' + g'/256 | ||
| 399 | add r7, r12, r7, lsr #8 @ | ||
| 400 | @ | ||
| 401 | @ This element is zero - use r14 @ | ||
| 402 | @ | ||
| 403 | add r3, r3, r14 @ b = r3 + delta | ||
| 404 | add r11, r11, r14, lsl #1 @ r = r11 + delta*2 | ||
| 405 | add r7, r7, r14, lsr #1 @ g = r7 + delta/2 | ||
| 406 | @ | ||
| 407 | orr r12, r3, r11, asr #1 @ check if clamping is needed... | ||
| 408 | orr r12, r12, r7 @ ...at all | ||
| 409 | movs r12, r12, asr #15 @ | ||
| 410 | beq 15f @ no clamp @ | ||
| 411 | movs r12, r3, asr #15 @ clamp b | ||
| 412 | mvnne r3, r12, lsr #15 @ | ||
| 413 | andne r3, r3, #0x7c00 @ mask b only if clamped | ||
| 414 | movs r12, r11, asr #16 @ clamp r | ||
| 415 | mvnne r11, r12, lsr #16 @ | ||
| 416 | movs r12, r7, asr #15 @ clamp g | ||
| 417 | mvnne r7, r12, lsr #15 @ | ||
| 418 | 15: @ no clamp @ | ||
| 419 | @ | ||
| 420 | ldrb r12, [r4], #1 @ r12 = Y' = *(Y'_p++) | ||
| 421 | @ | ||
| 422 | and r11, r11, #0xf800 @ pack pixel | ||
| 423 | and r7, r7, #0x7e00 @ r3 = pixel = (r & 0xf800) | | ||
| 424 | orr r11, r11, r7, lsr #4 @ ((g & 0x7e00) >> 4) | | ||
| 425 | orr r3, r11, r3, lsr #10 @ (b >> 10) | ||
| 426 | @ | ||
| 427 | movs r7, r3, lsr #8 @ store pixel | ||
| 428 | 20: @ | ||
| 429 | ldr r11, [r0] @ | ||
| 430 | tst r11, #LCD1_BUSY_MASK @ | ||
| 431 | bne 20b @ | ||
| 432 | str r7, [r0, #0x10] @ | ||
| 433 | 25: @ | ||
| 434 | ldr r11, [r0] @ | ||
| 435 | tst r11, #LCD1_BUSY_MASK @ | ||
| 436 | bne 25b @ | ||
| 437 | str r3, [r0, #0x10] @ | ||
| 438 | @ | ||
| 439 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 440 | add r12, r7, r7, asl #2 @ | ||
| 441 | add r12, r12, r12, asl #4 @ | ||
| 442 | add r7, r12, r7, asl #6 @ | ||
| 443 | @ compute R, G, and B | ||
| 444 | add r3, r8, r7 @ r3 = b' = Y + bu | ||
| 445 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 446 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 447 | @ | ||
| 448 | @ r8 = bu, r9 = rv, r10 = guv | ||
| 449 | @ | ||
| 450 | sub r12, r3, r3, lsr #5 @ r3 = 31/32*b' + b'/256 | ||
| 451 | add r3, r12, r3, lsr #8 @ | ||
| 452 | @ | ||
| 453 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r' + r'/256 | ||
| 454 | add r11, r12, r11, lsr #8 @ | ||
| 455 | @ | ||
| 456 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g' + g'/256 | ||
| 457 | add r7, r12, r7, lsr #8 @ | ||
| 458 | @ | ||
| 459 | add r12, r14, #0x100 @ | ||
| 460 | @ | ||
| 461 | add r3, r3, r12 @ b = r3 + delta | ||
| 462 | add r11, r11, r12, lsl #1 @ r = r11 + delta*2 | ||
| 463 | add r7, r7, r12, lsr #1 @ g = r7 + delta/2 | ||
| 464 | @ | ||
| 465 | orr r12, r3, r11, asr #1 @ check if clamping is needed... | ||
| 466 | orr r12, r12, r7 @ ...at all | ||
| 467 | movs r12, r12, asr #15 @ | ||
| 468 | beq 15f @ no clamp @ | ||
| 469 | movs r12, r3, asr #15 @ clamp b | ||
| 470 | mvnne r3, r12, lsr #15 @ | ||
| 471 | andne r3, r3, #0x7c00 @ mask b only if clamped | ||
| 472 | movs r12, r11, asr #16 @ clamp r | ||
| 473 | mvnne r11, r12, lsr #16 @ | ||
| 474 | movs r12, r7, asr #15 @ clamp g | ||
| 475 | mvnne r7, r12, lsr #15 @ | ||
| 476 | 15: @ no clamp @ | ||
| 477 | @ | ||
| 478 | ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride) | ||
| 479 | @ | ||
| 480 | and r11, r11, #0xf800 @ pack pixel | ||
| 481 | and r7, r7, #0x7e00 @ r3 = pixel = (r & 0xf800) | | ||
| 482 | orr r11, r11, r7, lsr #4 @ ((g & 0x7e00) >> 4) | | ||
| 483 | orr r3, r11, r3, lsr #10 @ (b >> 10) | ||
| 484 | @ | ||
| 485 | movs r7, r3, lsr #8 @ store pixel | ||
| 486 | 20: @ | ||
| 487 | ldr r11, [r0] @ | ||
| 488 | tst r11, #LCD1_BUSY_MASK @ | ||
| 489 | bne 20b @ | ||
| 490 | str r7, [r0, #0x10] @ | ||
| 491 | 25: @ | ||
| 492 | ldr r11, [r0] @ | ||
| 493 | tst r11, #LCD1_BUSY_MASK @ | ||
| 494 | bne 25b @ | ||
| 495 | str r3, [r0, #0x10] @ | ||
| 496 | @ | ||
| 497 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 498 | add r12, r7, r7, asl #2 @ | ||
| 499 | add r12, r12, r12, asl #4 @ | ||
| 500 | add r7, r12, r7, asl #6 @ | ||
| 501 | @ compute R, G, and B | ||
| 502 | add r3, r8, r7 @ r3 = b' = Y + bu | ||
| 503 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 504 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 505 | @ | ||
| 506 | sub r12, r3, r3, lsr #5 @ r3 = 31/32*b + b/256 | ||
| 507 | add r3, r12, r3, lsr #8 @ | ||
| 508 | @ | ||
| 509 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r + r/256 | ||
| 510 | add r11, r12, r11, lsr #8 @ | ||
| 511 | @ | ||
| 512 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g + g/256 | ||
| 513 | add r7, r12, r7, lsr #8 @ | ||
| 514 | @ | ||
| 515 | add r12, r14, #0x300 @ | ||
| 516 | @ | ||
| 517 | add r3, r3, r12 @ b = r3 + delta | ||
| 518 | add r11, r11, r12, lsl #1 @ r = r11 + delta*2 | ||
| 519 | add r7, r7, r12, lsr #1 @ g = r7 + delta/2 | ||
| 520 | @ | ||
| 521 | orr r12, r3, r11, asr #1 @ check if clamping is needed... | ||
| 522 | orr r12, r12, r7 @ ...at all | ||
| 523 | movs r12, r12, asr #15 @ | ||
| 524 | beq 15f @ no clamp @ | ||
| 525 | movs r12, r3, asr #15 @ clamp b | ||
| 526 | mvnne r3, r12, lsr #15 @ | ||
| 527 | andne r3, r3, #0x7c00 @ mask b only if clamped | ||
| 528 | movs r12, r11, asr #16 @ clamp r | ||
| 529 | mvnne r11, r12, lsr #16 @ | ||
| 530 | movs r12, r7, asr #15 @ clamp g | ||
| 531 | mvnne r7, r12, lsr #15 @ | ||
| 532 | 15: @ no clamp @ | ||
| 533 | @ | ||
| 534 | and r11, r11, #0xf800 @ pack pixel | ||
| 535 | and r7, r7, #0x7e00 @ r3 = pixel = (r & 0xf800) | | ||
| 536 | orr r11, r11, r7, lsr #4 @ ((g & 0x7e00) >> 4) | | ||
| 537 | orr r3, r11, r3, lsr #10 @ (b >> 10) | ||
| 538 | @ | ||
| 539 | movs r7, r3, lsr #8 @ store pixel | ||
| 540 | 20: @ | ||
| 541 | ldr r11, [r0] @ | ||
| 542 | tst r11, #LCD1_BUSY_MASK @ | ||
| 543 | bne 20b @ | ||
| 544 | str r7, [r0, #0x10] @ | ||
| 545 | 25: @ | ||
| 546 | ldr r11, [r0] @ | ||
| 547 | tst r11, #LCD1_BUSY_MASK @ | ||
| 548 | bne 25b @ | ||
| 549 | str r3, [r0, #0x10] @ | ||
| 550 | @ | ||
| 551 | subs r1, r1, #2 @ subtract block from width | ||
| 552 | bgt 10b @ loop line @ | ||
| 553 | @ | ||
| 554 | ldmpc regs=r4-r11 @ restore registers and return | ||
| 555 | .ltorg @ dump constant pool | ||
| 556 | .size lcd_write_yuv420_lines_odither, .-lcd_write_yuv420_lines_odither | ||