Mercurial > cgi-bin > hgwebdir.cgi > PR > Applications > VSs > VSs__H264__App

annotate libavcodec/arm/dsputil_arm.S @ 3:0b056460c67d

find changesets by author, revision, files, or words in the commit message
changed code to use VSs
author Nina Engelhardt <nengel@mailbox.tu-berlin.de>
date Mon, 29 Oct 2012 16:44:27 +0100
parents
children
rev   line source
nengel@2 1 @
nengel@2 2 @ ARMv4 optimized DSP utils
nengel@2 3 @ Copyright (c) 2004 AGAWA Koji <i (AT) atty (DOT) jp>
nengel@2 4 @
nengel@2 5 @ This file is part of FFmpeg.
nengel@2 6 @
nengel@2 7 @ FFmpeg is free software; you can redistribute it and/or
nengel@2 8 @ modify it under the terms of the GNU Lesser General Public
nengel@2 9 @ License as published by the Free Software Foundation; either
nengel@2 10 @ version 2.1 of the License, or (at your option) any later version.
nengel@2 11 @
nengel@2 12 @ FFmpeg is distributed in the hope that it will be useful,
nengel@2 13 @ but WITHOUT ANY WARRANTY; without even the implied warranty of
nengel@2 14 @ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
nengel@2 15 @ Lesser General Public License for more details.
nengel@2 16 @
nengel@2 17 @ You should have received a copy of the GNU Lesser General Public
nengel@2 18 @ License along with FFmpeg; if not, write to the Free Software
nengel@2 19 @ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
nengel@2 20 @
nengel@2 21
nengel@2 22 #include "config.h"
nengel@2 23 #include "asm.S"
nengel@2 24
nengel@2 25 preserve8
nengel@2 26
nengel@2 27 #if !HAVE_PLD
nengel@2 28 .macro pld reg
nengel@2 29 .endm
nengel@2 30 #endif
nengel@2 31
nengel@2 32 #if HAVE_ARMV5TE
nengel@2 33 function ff_prefetch_arm, export=1
nengel@2 34 subs r2, r2, #1
nengel@2 35 pld [r0]
nengel@2 36 add r0, r0, r1
nengel@2 37 bne ff_prefetch_arm
nengel@2 38 bx lr
nengel@2 39 endfunc
nengel@2 40 #endif
nengel@2 41
nengel@2 42 .macro ALIGN_QWORD_D shift, Rd0, Rd1, Rd2, Rd3, Rn0, Rn1, Rn2, Rn3, Rn4
nengel@2 43 mov \Rd0, \Rn0, lsr #(\shift * 8)
nengel@2 44 mov \Rd1, \Rn1, lsr #(\shift * 8)
nengel@2 45 mov \Rd2, \Rn2, lsr #(\shift * 8)
nengel@2 46 mov \Rd3, \Rn3, lsr #(\shift * 8)
nengel@2 47 orr \Rd0, \Rd0, \Rn1, lsl #(32 - \shift * 8)
nengel@2 48 orr \Rd1, \Rd1, \Rn2, lsl #(32 - \shift * 8)
nengel@2 49 orr \Rd2, \Rd2, \Rn3, lsl #(32 - \shift * 8)
nengel@2 50 orr \Rd3, \Rd3, \Rn4, lsl #(32 - \shift * 8)
nengel@2 51 .endm
nengel@2 52 .macro ALIGN_DWORD shift, R0, R1, R2
nengel@2 53 mov \R0, \R0, lsr #(\shift * 8)
nengel@2 54 orr \R0, \R0, \R1, lsl #(32 - \shift * 8)
nengel@2 55 mov \R1, \R1, lsr #(\shift * 8)
nengel@2 56 orr \R1, \R1, \R2, lsl #(32 - \shift * 8)
nengel@2 57 .endm
nengel@2 58 .macro ALIGN_DWORD_D shift, Rdst0, Rdst1, Rsrc0, Rsrc1, Rsrc2
nengel@2 59 mov \Rdst0, \Rsrc0, lsr #(\shift * 8)
nengel@2 60 mov \Rdst1, \Rsrc1, lsr #(\shift * 8)
nengel@2 61 orr \Rdst0, \Rdst0, \Rsrc1, lsl #(32 - (\shift * 8))
nengel@2 62 orr \Rdst1, \Rdst1, \Rsrc2, lsl #(32 - (\shift * 8))
nengel@2 63 .endm
nengel@2 64
nengel@2 65 .macro RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask
nengel@2 66 @ Rd = (Rn | Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1)
nengel@2 67 @ Rmask = 0xFEFEFEFE
nengel@2 68 @ Rn = destroy
nengel@2 69 eor \Rd0, \Rn0, \Rm0
nengel@2 70 eor \Rd1, \Rn1, \Rm1
nengel@2 71 orr \Rn0, \Rn0, \Rm0
nengel@2 72 orr \Rn1, \Rn1, \Rm1
nengel@2 73 and \Rd0, \Rd0, \Rmask
nengel@2 74 and \Rd1, \Rd1, \Rmask
nengel@2 75 sub \Rd0, \Rn0, \Rd0, lsr #1
nengel@2 76 sub \Rd1, \Rn1, \Rd1, lsr #1
nengel@2 77 .endm
nengel@2 78
nengel@2 79 .macro NO_RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask
nengel@2 80 @ Rd = (Rn & Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1)
nengel@2 81 @ Rmask = 0xFEFEFEFE
nengel@2 82 @ Rn = destroy
nengel@2 83 eor \Rd0, \Rn0, \Rm0
nengel@2 84 eor \Rd1, \Rn1, \Rm1
nengel@2 85 and \Rn0, \Rn0, \Rm0
nengel@2 86 and \Rn1, \Rn1, \Rm1
nengel@2 87 and \Rd0, \Rd0, \Rmask
nengel@2 88 and \Rd1, \Rd1, \Rmask
nengel@2 89 add \Rd0, \Rn0, \Rd0, lsr #1
nengel@2 90 add \Rd1, \Rn1, \Rd1, lsr #1
nengel@2 91 .endm
nengel@2 92
nengel@2 93 .macro JMP_ALIGN tmp, reg
nengel@2 94 ands \tmp, \reg, #3
nengel@2 95 bic \reg, \reg, #3
nengel@2 96 beq 1f
nengel@2 97 subs \tmp, \tmp, #1
nengel@2 98 beq 2f
nengel@2 99 subs \tmp, \tmp, #1
nengel@2 100 beq 3f
nengel@2 101 b 4f
nengel@2 102 .endm
nengel@2 103
nengel@2 104 @ ----------------------------------------------------------------
nengel@2 105 .align 5
nengel@2 106 function ff_put_pixels16_arm, export=1
nengel@2 107 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 108 @ block = word aligned, pixles = unaligned
nengel@2 109 pld [r1]
nengel@2 110 push {r4-r11, lr}
nengel@2 111 JMP_ALIGN r5, r1
nengel@2 112 1:
nengel@2 113 ldm r1, {r4-r7}
nengel@2 114 add r1, r1, r2
nengel@2 115 stm r0, {r4-r7}
nengel@2 116 pld [r1]
nengel@2 117 subs r3, r3, #1
nengel@2 118 add r0, r0, r2
nengel@2 119 bne 1b
nengel@2 120 pop {r4-r11, pc}
nengel@2 121 .align 5
nengel@2 122 2:
nengel@2 123 ldm r1, {r4-r8}
nengel@2 124 add r1, r1, r2
nengel@2 125 ALIGN_QWORD_D 1, r9, r10, r11, r12, r4, r5, r6, r7, r8
nengel@2 126 pld [r1]
nengel@2 127 subs r3, r3, #1
nengel@2 128 stm r0, {r9-r12}
nengel@2 129 add r0, r0, r2
nengel@2 130 bne 2b
nengel@2 131 pop {r4-r11, pc}
nengel@2 132 .align 5
nengel@2 133 3:
nengel@2 134 ldm r1, {r4-r8}
nengel@2 135 add r1, r1, r2
nengel@2 136 ALIGN_QWORD_D 2, r9, r10, r11, r12, r4, r5, r6, r7, r8
nengel@2 137 pld [r1]
nengel@2 138 subs r3, r3, #1
nengel@2 139 stm r0, {r9-r12}
nengel@2 140 add r0, r0, r2
nengel@2 141 bne 3b
nengel@2 142 pop {r4-r11, pc}
nengel@2 143 .align 5
nengel@2 144 4:
nengel@2 145 ldm r1, {r4-r8}
nengel@2 146 add r1, r1, r2
nengel@2 147 ALIGN_QWORD_D 3, r9, r10, r11, r12, r4, r5, r6, r7, r8
nengel@2 148 pld [r1]
nengel@2 149 subs r3, r3, #1
nengel@2 150 stm r0, {r9-r12}
nengel@2 151 add r0, r0, r2
nengel@2 152 bne 4b
nengel@2 153 pop {r4-r11,pc}
nengel@2 154 endfunc
nengel@2 155
nengel@2 156 @ ----------------------------------------------------------------
nengel@2 157 .align 5
nengel@2 158 function ff_put_pixels8_arm, export=1
nengel@2 159 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 160 @ block = word aligned, pixles = unaligned
nengel@2 161 pld [r1]
nengel@2 162 push {r4-r5,lr}
nengel@2 163 JMP_ALIGN r5, r1
nengel@2 164 1:
nengel@2 165 ldm r1, {r4-r5}
nengel@2 166 add r1, r1, r2
nengel@2 167 subs r3, r3, #1
nengel@2 168 pld [r1]
nengel@2 169 stm r0, {r4-r5}
nengel@2 170 add r0, r0, r2
nengel@2 171 bne 1b
nengel@2 172 pop {r4-r5,pc}
nengel@2 173 .align 5
nengel@2 174 2:
nengel@2 175 ldm r1, {r4-r5, r12}
nengel@2 176 add r1, r1, r2
nengel@2 177 ALIGN_DWORD 1, r4, r5, r12
nengel@2 178 pld [r1]
nengel@2 179 subs r3, r3, #1
nengel@2 180 stm r0, {r4-r5}
nengel@2 181 add r0, r0, r2
nengel@2 182 bne 2b
nengel@2 183 pop {r4-r5,pc}
nengel@2 184 .align 5
nengel@2 185 3:
nengel@2 186 ldm r1, {r4-r5, r12}
nengel@2 187 add r1, r1, r2
nengel@2 188 ALIGN_DWORD 2, r4, r5, r12
nengel@2 189 pld [r1]
nengel@2 190 subs r3, r3, #1
nengel@2 191 stm r0, {r4-r5}
nengel@2 192 add r0, r0, r2
nengel@2 193 bne 3b
nengel@2 194 pop {r4-r5,pc}
nengel@2 195 .align 5
nengel@2 196 4:
nengel@2 197 ldm r1, {r4-r5, r12}
nengel@2 198 add r1, r1, r2
nengel@2 199 ALIGN_DWORD 3, r4, r5, r12
nengel@2 200 pld [r1]
nengel@2 201 subs r3, r3, #1
nengel@2 202 stm r0, {r4-r5}
nengel@2 203 add r0, r0, r2
nengel@2 204 bne 4b
nengel@2 205 pop {r4-r5,pc}
nengel@2 206 endfunc
nengel@2 207
nengel@2 208 @ ----------------------------------------------------------------
nengel@2 209 .align 5
nengel@2 210 function ff_put_pixels8_x2_arm, export=1
nengel@2 211 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 212 @ block = word aligned, pixles = unaligned
nengel@2 213 pld [r1]
nengel@2 214 push {r4-r10,lr}
nengel@2 215 ldr r12, =0xfefefefe
nengel@2 216 JMP_ALIGN r5, r1
nengel@2 217 1:
nengel@2 218 ldm r1, {r4-r5, r10}
nengel@2 219 add r1, r1, r2
nengel@2 220 ALIGN_DWORD_D 1, r6, r7, r4, r5, r10
nengel@2 221 pld [r1]
nengel@2 222 RND_AVG32 r8, r9, r4, r5, r6, r7, r12
nengel@2 223 subs r3, r3, #1
nengel@2 224 stm r0, {r8-r9}
nengel@2 225 add r0, r0, r2
nengel@2 226 bne 1b
nengel@2 227 pop {r4-r10,pc}
nengel@2 228 .align 5
nengel@2 229 2:
nengel@2 230 ldm r1, {r4-r5, r10}
nengel@2 231 add r1, r1, r2
nengel@2 232 ALIGN_DWORD_D 1, r6, r7, r4, r5, r10
nengel@2 233 ALIGN_DWORD_D 2, r8, r9, r4, r5, r10
nengel@2 234 pld [r1]
nengel@2 235 RND_AVG32 r4, r5, r6, r7, r8, r9, r12
nengel@2 236 subs r3, r3, #1
nengel@2 237 stm r0, {r4-r5}
nengel@2 238 add r0, r0, r2
nengel@2 239 bne 2b
nengel@2 240 pop {r4-r10,pc}
nengel@2 241 .align 5
nengel@2 242 3:
nengel@2 243 ldm r1, {r4-r5, r10}
nengel@2 244 add r1, r1, r2
nengel@2 245 ALIGN_DWORD_D 2, r6, r7, r4, r5, r10
nengel@2 246 ALIGN_DWORD_D 3, r8, r9, r4, r5, r10
nengel@2 247 pld [r1]
nengel@2 248 RND_AVG32 r4, r5, r6, r7, r8, r9, r12
nengel@2 249 subs r3, r3, #1
nengel@2 250 stm r0, {r4-r5}
nengel@2 251 add r0, r0, r2
nengel@2 252 bne 3b
nengel@2 253 pop {r4-r10,pc}
nengel@2 254 .align 5
nengel@2 255 4:
nengel@2 256 ldm r1, {r4-r5, r10}
nengel@2 257 add r1, r1, r2
nengel@2 258 ALIGN_DWORD_D 3, r6, r7, r4, r5, r10
nengel@2 259 pld [r1]
nengel@2 260 RND_AVG32 r8, r9, r6, r7, r5, r10, r12
nengel@2 261 subs r3, r3, #1
nengel@2 262 stm r0, {r8-r9}
nengel@2 263 add r0, r0, r2
nengel@2 264 bne 4b
nengel@2 265 pop {r4-r10,pc}
nengel@2 266 endfunc
nengel@2 267
nengel@2 268 .align 5
nengel@2 269 function ff_put_no_rnd_pixels8_x2_arm, export=1
nengel@2 270 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 271 @ block = word aligned, pixles = unaligned
nengel@2 272 pld [r1]
nengel@2 273 push {r4-r10,lr}
nengel@2 274 ldr r12, =0xfefefefe
nengel@2 275 JMP_ALIGN r5, r1
nengel@2 276 1:
nengel@2 277 ldm r1, {r4-r5, r10}
nengel@2 278 add r1, r1, r2
nengel@2 279 ALIGN_DWORD_D 1, r6, r7, r4, r5, r10
nengel@2 280 pld [r1]
nengel@2 281 NO_RND_AVG32 r8, r9, r4, r5, r6, r7, r12
nengel@2 282 subs r3, r3, #1
nengel@2 283 stm r0, {r8-r9}
nengel@2 284 add r0, r0, r2
nengel@2 285 bne 1b
nengel@2 286 pop {r4-r10,pc}
nengel@2 287 .align 5
nengel@2 288 2:
nengel@2 289 ldm r1, {r4-r5, r10}
nengel@2 290 add r1, r1, r2
nengel@2 291 ALIGN_DWORD_D 1, r6, r7, r4, r5, r10
nengel@2 292 ALIGN_DWORD_D 2, r8, r9, r4, r5, r10
nengel@2 293 pld [r1]
nengel@2 294 NO_RND_AVG32 r4, r5, r6, r7, r8, r9, r12
nengel@2 295 subs r3, r3, #1
nengel@2 296 stm r0, {r4-r5}
nengel@2 297 add r0, r0, r2
nengel@2 298 bne 2b
nengel@2 299 pop {r4-r10,pc}
nengel@2 300 .align 5
nengel@2 301 3:
nengel@2 302 ldm r1, {r4-r5, r10}
nengel@2 303 add r1, r1, r2
nengel@2 304 ALIGN_DWORD_D 2, r6, r7, r4, r5, r10
nengel@2 305 ALIGN_DWORD_D 3, r8, r9, r4, r5, r10
nengel@2 306 pld [r1]
nengel@2 307 NO_RND_AVG32 r4, r5, r6, r7, r8, r9, r12
nengel@2 308 subs r3, r3, #1
nengel@2 309 stm r0, {r4-r5}
nengel@2 310 add r0, r0, r2
nengel@2 311 bne 3b
nengel@2 312 pop {r4-r10,pc}
nengel@2 313 .align 5
nengel@2 314 4:
nengel@2 315 ldm r1, {r4-r5, r10}
nengel@2 316 add r1, r1, r2
nengel@2 317 ALIGN_DWORD_D 3, r6, r7, r4, r5, r10
nengel@2 318 pld [r1]
nengel@2 319 NO_RND_AVG32 r8, r9, r6, r7, r5, r10, r12
nengel@2 320 subs r3, r3, #1
nengel@2 321 stm r0, {r8-r9}
nengel@2 322 add r0, r0, r2
nengel@2 323 bne 4b
nengel@2 324 pop {r4-r10,pc}
nengel@2 325 endfunc
nengel@2 326
nengel@2 327
nengel@2 328 @ ----------------------------------------------------------------
nengel@2 329 .align 5
nengel@2 330 function ff_put_pixels8_y2_arm, export=1
nengel@2 331 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 332 @ block = word aligned, pixles = unaligned
nengel@2 333 pld [r1]
nengel@2 334 push {r4-r11,lr}
nengel@2 335 mov r3, r3, lsr #1
nengel@2 336 ldr r12, =0xfefefefe
nengel@2 337 JMP_ALIGN r5, r1
nengel@2 338 1:
nengel@2 339 ldm r1, {r4-r5}
nengel@2 340 add r1, r1, r2
nengel@2 341 6: ldm r1, {r6-r7}
nengel@2 342 add r1, r1, r2
nengel@2 343 pld [r1]
nengel@2 344 RND_AVG32 r8, r9, r4, r5, r6, r7, r12
nengel@2 345 ldm r1, {r4-r5}
nengel@2 346 add r1, r1, r2
nengel@2 347 stm r0, {r8-r9}
nengel@2 348 add r0, r0, r2
nengel@2 349 pld [r1]
nengel@2 350 RND_AVG32 r8, r9, r6, r7, r4, r5, r12
nengel@2 351 subs r3, r3, #1
nengel@2 352 stm r0, {r8-r9}
nengel@2 353 add r0, r0, r2
nengel@2 354 bne 6b
nengel@2 355 pop {r4-r11,pc}
nengel@2 356 .align 5
nengel@2 357 2:
nengel@2 358 ldm r1, {r4-r6}
nengel@2 359 add r1, r1, r2
nengel@2 360 pld [r1]
nengel@2 361 ALIGN_DWORD 1, r4, r5, r6
nengel@2 362 6: ldm r1, {r7-r9}
nengel@2 363 add r1, r1, r2
nengel@2 364 pld [r1]
nengel@2 365 ALIGN_DWORD 1, r7, r8, r9
nengel@2 366 RND_AVG32 r10, r11, r4, r5, r7, r8, r12
nengel@2 367 stm r0, {r10-r11}
nengel@2 368 add r0, r0, r2
nengel@2 369 ldm r1, {r4-r6}
nengel@2 370 add r1, r1, r2
nengel@2 371 pld [r1]
nengel@2 372 ALIGN_DWORD 1, r4, r5, r6
nengel@2 373 subs r3, r3, #1
nengel@2 374 RND_AVG32 r10, r11, r7, r8, r4, r5, r12
nengel@2 375 stm r0, {r10-r11}
nengel@2 376 add r0, r0, r2
nengel@2 377 bne 6b
nengel@2 378 pop {r4-r11,pc}
nengel@2 379 .align 5
nengel@2 380 3:
nengel@2 381 ldm r1, {r4-r6}
nengel@2 382 add r1, r1, r2
nengel@2 383 pld [r1]
nengel@2 384 ALIGN_DWORD 2, r4, r5, r6
nengel@2 385 6: ldm r1, {r7-r9}
nengel@2 386 add r1, r1, r2
nengel@2 387 pld [r1]
nengel@2 388 ALIGN_DWORD 2, r7, r8, r9
nengel@2 389 RND_AVG32 r10, r11, r4, r5, r7, r8, r12
nengel@2 390 stm r0, {r10-r11}
nengel@2 391 add r0, r0, r2
nengel@2 392 ldm r1, {r4-r6}
nengel@2 393 add r1, r1, r2
nengel@2 394 pld [r1]
nengel@2 395 ALIGN_DWORD 2, r4, r5, r6
nengel@2 396 subs r3, r3, #1
nengel@2 397 RND_AVG32 r10, r11, r7, r8, r4, r5, r12
nengel@2 398 stm r0, {r10-r11}
nengel@2 399 add r0, r0, r2
nengel@2 400 bne 6b
nengel@2 401 pop {r4-r11,pc}
nengel@2 402 .align 5
nengel@2 403 4:
nengel@2 404 ldm r1, {r4-r6}
nengel@2 405 add r1, r1, r2
nengel@2 406 pld [r1]
nengel@2 407 ALIGN_DWORD 3, r4, r5, r6
nengel@2 408 6: ldm r1, {r7-r9}
nengel@2 409 add r1, r1, r2
nengel@2 410 pld [r1]
nengel@2 411 ALIGN_DWORD 3, r7, r8, r9
nengel@2 412 RND_AVG32 r10, r11, r4, r5, r7, r8, r12
nengel@2 413 stm r0, {r10-r11}
nengel@2 414 add r0, r0, r2
nengel@2 415 ldm r1, {r4-r6}
nengel@2 416 add r1, r1, r2
nengel@2 417 pld [r1]
nengel@2 418 ALIGN_DWORD 3, r4, r5, r6
nengel@2 419 subs r3, r3, #1
nengel@2 420 RND_AVG32 r10, r11, r7, r8, r4, r5, r12
nengel@2 421 stm r0, {r10-r11}
nengel@2 422 add r0, r0, r2
nengel@2 423 bne 6b
nengel@2 424 pop {r4-r11,pc}
nengel@2 425 endfunc
nengel@2 426
nengel@2 427 .align 5
nengel@2 428 function ff_put_no_rnd_pixels8_y2_arm, export=1
nengel@2 429 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 430 @ block = word aligned, pixles = unaligned
nengel@2 431 pld [r1]
nengel@2 432 push {r4-r11,lr}
nengel@2 433 mov r3, r3, lsr #1
nengel@2 434 ldr r12, =0xfefefefe
nengel@2 435 JMP_ALIGN r5, r1
nengel@2 436 1:
nengel@2 437 ldm r1, {r4-r5}
nengel@2 438 add r1, r1, r2
nengel@2 439 6: ldm r1, {r6-r7}
nengel@2 440 add r1, r1, r2
nengel@2 441 pld [r1]
nengel@2 442 NO_RND_AVG32 r8, r9, r4, r5, r6, r7, r12
nengel@2 443 ldm r1, {r4-r5}
nengel@2 444 add r1, r1, r2
nengel@2 445 stm r0, {r8-r9}
nengel@2 446 add r0, r0, r2
nengel@2 447 pld [r1]
nengel@2 448 NO_RND_AVG32 r8, r9, r6, r7, r4, r5, r12
nengel@2 449 subs r3, r3, #1
nengel@2 450 stm r0, {r8-r9}
nengel@2 451 add r0, r0, r2
nengel@2 452 bne 6b
nengel@2 453 pop {r4-r11,pc}
nengel@2 454 .align 5
nengel@2 455 2:
nengel@2 456 ldm r1, {r4-r6}
nengel@2 457 add r1, r1, r2
nengel@2 458 pld [r1]
nengel@2 459 ALIGN_DWORD 1, r4, r5, r6
nengel@2 460 6: ldm r1, {r7-r9}
nengel@2 461 add r1, r1, r2
nengel@2 462 pld [r1]
nengel@2 463 ALIGN_DWORD 1, r7, r8, r9
nengel@2 464 NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12
nengel@2 465 stm r0, {r10-r11}
nengel@2 466 add r0, r0, r2
nengel@2 467 ldm r1, {r4-r6}
nengel@2 468 add r1, r1, r2
nengel@2 469 pld [r1]
nengel@2 470 ALIGN_DWORD 1, r4, r5, r6
nengel@2 471 subs r3, r3, #1
nengel@2 472 NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12
nengel@2 473 stm r0, {r10-r11}
nengel@2 474 add r0, r0, r2
nengel@2 475 bne 6b
nengel@2 476 pop {r4-r11,pc}
nengel@2 477 .align 5
nengel@2 478 3:
nengel@2 479 ldm r1, {r4-r6}
nengel@2 480 add r1, r1, r2
nengel@2 481 pld [r1]
nengel@2 482 ALIGN_DWORD 2, r4, r5, r6
nengel@2 483 6: ldm r1, {r7-r9}
nengel@2 484 add r1, r1, r2
nengel@2 485 pld [r1]
nengel@2 486 ALIGN_DWORD 2, r7, r8, r9
nengel@2 487 NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12
nengel@2 488 stm r0, {r10-r11}
nengel@2 489 add r0, r0, r2
nengel@2 490 ldm r1, {r4-r6}
nengel@2 491 add r1, r1, r2
nengel@2 492 pld [r1]
nengel@2 493 ALIGN_DWORD 2, r4, r5, r6
nengel@2 494 subs r3, r3, #1
nengel@2 495 NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12
nengel@2 496 stm r0, {r10-r11}
nengel@2 497 add r0, r0, r2
nengel@2 498 bne 6b
nengel@2 499 pop {r4-r11,pc}
nengel@2 500 .align 5
nengel@2 501 4:
nengel@2 502 ldm r1, {r4-r6}
nengel@2 503 add r1, r1, r2
nengel@2 504 pld [r1]
nengel@2 505 ALIGN_DWORD 3, r4, r5, r6
nengel@2 506 6: ldm r1, {r7-r9}
nengel@2 507 add r1, r1, r2
nengel@2 508 pld [r1]
nengel@2 509 ALIGN_DWORD 3, r7, r8, r9
nengel@2 510 NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12
nengel@2 511 stm r0, {r10-r11}
nengel@2 512 add r0, r0, r2
nengel@2 513 ldm r1, {r4-r6}
nengel@2 514 add r1, r1, r2
nengel@2 515 pld [r1]
nengel@2 516 ALIGN_DWORD 3, r4, r5, r6
nengel@2 517 subs r3, r3, #1
nengel@2 518 NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12
nengel@2 519 stm r0, {r10-r11}
nengel@2 520 add r0, r0, r2
nengel@2 521 bne 6b
nengel@2 522 pop {r4-r11,pc}
nengel@2 523 endfunc
nengel@2 524
nengel@2 525 .ltorg
nengel@2 526
nengel@2 527 @ ----------------------------------------------------------------
nengel@2 528 .macro RND_XY2_IT align, rnd
nengel@2 529 @ l1= (a & 0x03030303) + (b & 0x03030303) ?(+ 0x02020202)
nengel@2 530 @ h1= ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2)
nengel@2 531 .if \align == 0
nengel@2 532 ldm r1, {r6-r8}
nengel@2 533 .elseif \align == 3
nengel@2 534 ldm r1, {r5-r7}
nengel@2 535 .else
nengel@2 536 ldm r1, {r8-r10}
nengel@2 537 .endif
nengel@2 538 add r1, r1, r2
nengel@2 539 pld [r1]
nengel@2 540 .if \align == 0
nengel@2 541 ALIGN_DWORD_D 1, r4, r5, r6, r7, r8
nengel@2 542 .elseif \align == 1
nengel@2 543 ALIGN_DWORD_D 1, r4, r5, r8, r9, r10
nengel@2 544 ALIGN_DWORD_D 2, r6, r7, r8, r9, r10
nengel@2 545 .elseif \align == 2
nengel@2 546 ALIGN_DWORD_D 2, r4, r5, r8, r9, r10
nengel@2 547 ALIGN_DWORD_D 3, r6, r7, r8, r9, r10
nengel@2 548 .elseif \align == 3
nengel@2 549 ALIGN_DWORD_D 3, r4, r5, r5, r6, r7
nengel@2 550 .endif
nengel@2 551 ldr r14, =0x03030303
nengel@2 552 tst r3, #1
nengel@2 553 and r8, r4, r14
nengel@2 554 and r9, r5, r14
nengel@2 555 and r10, r6, r14
nengel@2 556 and r11, r7, r14
nengel@2 557 andeq r14, r14, r14, \rnd #1
nengel@2 558 add r8, r8, r10
nengel@2 559 add r9, r9, r11
nengel@2 560 ldr r12, =0xfcfcfcfc >> 2
nengel@2 561 addeq r8, r8, r14
nengel@2 562 addeq r9, r9, r14
nengel@2 563 and r4, r12, r4, lsr #2
nengel@2 564 and r5, r12, r5, lsr #2
nengel@2 565 and r6, r12, r6, lsr #2
nengel@2 566 and r7, r12, r7, lsr #2
nengel@2 567 add r10, r4, r6
nengel@2 568 add r11, r5, r7
nengel@2 569 subs r3, r3, #1
nengel@2 570 .endm
nengel@2 571
nengel@2 572 .macro RND_XY2_EXPAND align, rnd
nengel@2 573 RND_XY2_IT \align, \rnd
nengel@2 574 6: push {r8-r11}
nengel@2 575 RND_XY2_IT \align, \rnd
nengel@2 576 pop {r4-r7}
nengel@2 577 add r4, r4, r8
nengel@2 578 add r5, r5, r9
nengel@2 579 ldr r14, =0x0f0f0f0f
nengel@2 580 add r6, r6, r10
nengel@2 581 add r7, r7, r11
nengel@2 582 and r4, r14, r4, lsr #2
nengel@2 583 and r5, r14, r5, lsr #2
nengel@2 584 add r4, r4, r6
nengel@2 585 add r5, r5, r7
nengel@2 586 stm r0, {r4-r5}
nengel@2 587 add r0, r0, r2
nengel@2 588 bge 6b
nengel@2 589 pop {r4-r11,pc}
nengel@2 590 .endm
nengel@2 591
nengel@2 592 .align 5
nengel@2 593 function ff_put_pixels8_xy2_arm, export=1
nengel@2 594 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 595 @ block = word aligned, pixles = unaligned
nengel@2 596 pld [r1]
nengel@2 597 push {r4-r11,lr} @ R14 is also called LR
nengel@2 598 JMP_ALIGN r5, r1
nengel@2 599 1: RND_XY2_EXPAND 0, lsl
nengel@2 600 .align 5
nengel@2 601 2: RND_XY2_EXPAND 1, lsl
nengel@2 602 .align 5
nengel@2 603 3: RND_XY2_EXPAND 2, lsl
nengel@2 604 .align 5
nengel@2 605 4: RND_XY2_EXPAND 3, lsl
nengel@2 606 endfunc
nengel@2 607
nengel@2 608 .align 5
nengel@2 609 function ff_put_no_rnd_pixels8_xy2_arm, export=1
nengel@2 610 @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
nengel@2 611 @ block = word aligned, pixles = unaligned
nengel@2 612 pld [r1]
nengel@2 613 push {r4-r11,lr}
nengel@2 614 JMP_ALIGN r5, r1
nengel@2 615 1: RND_XY2_EXPAND 0, lsr
nengel@2 616 .align 5
nengel@2 617 2: RND_XY2_EXPAND 1, lsr
nengel@2 618 .align 5
nengel@2 619 3: RND_XY2_EXPAND 2, lsr
nengel@2 620 .align 5
nengel@2 621 4: RND_XY2_EXPAND 3, lsr
nengel@2 622 endfunc
nengel@2 623
nengel@2 624 .align 5
nengel@2 625 @ void ff_add_pixels_clamped_arm(int16_t *block, uint8_t *dest, int stride)
nengel@2 626 function ff_add_pixels_clamped_arm, export=1
nengel@2 627 push {r4-r10}
nengel@2 628 mov r10, #8
nengel@2 629 1:
nengel@2 630 ldr r4, [r1] /* load dest */
nengel@2 631 /* block[0] and block[1]*/
nengel@2 632 ldrsh r5, [r0]
nengel@2 633 ldrsh r7, [r0, #2]
nengel@2 634 and r6, r4, #0xFF
nengel@2 635 and r8, r4, #0xFF00
nengel@2 636 add r6, r5, r6
nengel@2 637 add r8, r7, r8, lsr #8
nengel@2 638 mvn r5, r5
nengel@2 639 mvn r7, r7
nengel@2 640 tst r6, #0x100
nengel@2 641 movne r6, r5, lsr #24
nengel@2 642 tst r8, #0x100
nengel@2 643 movne r8, r7, lsr #24
nengel@2 644 mov r9, r6
nengel@2 645 ldrsh r5, [r0, #4] /* moved form [A] */
nengel@2 646 orr r9, r9, r8, lsl #8
nengel@2 647 /* block[2] and block[3] */
nengel@2 648 /* [A] */
nengel@2 649 ldrsh r7, [r0, #6]
nengel@2 650 and r6, r4, #0xFF0000
nengel@2 651 and r8, r4, #0xFF000000
nengel@2 652 add r6, r5, r6, lsr #16
nengel@2 653 add r8, r7, r8, lsr #24
nengel@2 654 mvn r5, r5
nengel@2 655 mvn r7, r7
nengel@2 656 tst r6, #0x100
nengel@2 657 movne r6, r5, lsr #24
nengel@2 658 tst r8, #0x100
nengel@2 659 movne r8, r7, lsr #24
nengel@2 660 orr r9, r9, r6, lsl #16
nengel@2 661 ldr r4, [r1, #4] /* moved form [B] */
nengel@2 662 orr r9, r9, r8, lsl #24
nengel@2 663 /* store dest */
nengel@2 664 ldrsh r5, [r0, #8] /* moved form [C] */
nengel@2 665 str r9, [r1]
nengel@2 666
nengel@2 667 /* load dest */
nengel@2 668 /* [B] */
nengel@2 669 /* block[4] and block[5] */
nengel@2 670 /* [C] */
nengel@2 671 ldrsh r7, [r0, #10]
nengel@2 672 and r6, r4, #0xFF
nengel@2 673 and r8, r4, #0xFF00
nengel@2 674 add r6, r5, r6
nengel@2 675 add r8, r7, r8, lsr #8
nengel@2 676 mvn r5, r5
nengel@2 677 mvn r7, r7
nengel@2 678 tst r6, #0x100
nengel@2 679 movne r6, r5, lsr #24
nengel@2 680 tst r8, #0x100
nengel@2 681 movne r8, r7, lsr #24
nengel@2 682 mov r9, r6
nengel@2 683 ldrsh r5, [r0, #12] /* moved from [D] */
nengel@2 684 orr r9, r9, r8, lsl #8
nengel@2 685 /* block[6] and block[7] */
nengel@2 686 /* [D] */
nengel@2 687 ldrsh r7, [r0, #14]
nengel@2 688 and r6, r4, #0xFF0000
nengel@2 689 and r8, r4, #0xFF000000
nengel@2 690 add r6, r5, r6, lsr #16
nengel@2 691 add r8, r7, r8, lsr #24
nengel@2 692 mvn r5, r5
nengel@2 693 mvn r7, r7
nengel@2 694 tst r6, #0x100
nengel@2 695 movne r6, r5, lsr #24
nengel@2 696 tst r8, #0x100
nengel@2 697 movne r8, r7, lsr #24
nengel@2 698 orr r9, r9, r6, lsl #16
nengel@2 699 add r0, r0, #16 /* moved from [E] */
nengel@2 700 orr r9, r9, r8, lsl #24
nengel@2 701 subs r10, r10, #1 /* moved from [F] */
nengel@2 702 /* store dest */
nengel@2 703 str r9, [r1, #4]
nengel@2 704
nengel@2 705 /* [E] */
nengel@2 706 /* [F] */
nengel@2 707 add r1, r1, r2
nengel@2 708 bne 1b
nengel@2 709
nengel@2 710 pop {r4-r10}
nengel@2 711 bx lr
nengel@2 712 endfunc