DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Mercurial (d8847129d134)

VCS Links

Line Code
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
/* NEON optimized code (C) COPYRIGHT 2009 Motorola
 *
 * Use of this source code is governed by a BSD-style license that can be
 * found in the LICENSE file.
 */

#include "SkBitmapProcState.h"
#include "SkPerspIter.h"
#include "SkShader.h"
#include "SkUtils.h"
#include "SkUtilsArm.h"
#include "SkBitmapProcState_utils.h"

/*  returns 0...(n-1) given any x (positive or negative).

    As an example, if n (which is always positive) is 5...

          x: -8 -7 -6 -5 -4 -3 -2 -1  0  1  2  3  4  5  6  7  8
    returns:  2  3  4  0  1  2  3  4  0  1  2  3  4  0  1  2  3
 */
static inline int sk_int_mod(int x, int n) {
    SkASSERT(n > 0);
    if ((unsigned)x >= (unsigned)n) {
        if (x < 0) {
            x = n + ~(~x % n);
        } else {
            x = x % n;
        }
    }
    return x;
}

void decal_nofilter_scale(uint32_t dst[], SkFixed fx, SkFixed dx, int count);
void decal_filter_scale(uint32_t dst[], SkFixed fx, SkFixed dx, int count);

#include "SkBitmapProcState_matrix_template.h"

///////////////////////////////////////////////////////////////////////////////

// Compile neon code paths if needed
#if !SK_ARM_NEON_IS_NONE

// These are defined in src/opts/SkBitmapProcState_matrixProcs_neon.cpp
extern const SkBitmapProcState::MatrixProc ClampX_ClampY_Procs_neon[];
extern const SkBitmapProcState::MatrixProc RepeatX_RepeatY_Procs_neon[];

#endif // !SK_ARM_NEON_IS_NONE

// Compile non-neon code path if needed
#if !SK_ARM_NEON_IS_ALWAYS
#define MAKENAME(suffix)        ClampX_ClampY ## suffix
#define TILEX_PROCF(fx, max)    SkClampMax((fx) >> 16, max)
#define TILEY_PROCF(fy, max)    SkClampMax((fy) >> 16, max)
#define TILEX_LOW_BITS(fx, max) (((fx) >> 12) & 0xF)
#define TILEY_LOW_BITS(fy, max) (((fy) >> 12) & 0xF)
#define CHECK_FOR_DECAL
#include "SkBitmapProcState_matrix.h"

struct ClampTileProcs {
    static unsigned X(const SkBitmapProcState&, SkFixed fx, int max) {
        return SkClampMax(fx >> 16, max);
    }
    static unsigned Y(const SkBitmapProcState&, SkFixed fy, int max) {
        return SkClampMax(fy >> 16, max);
    }
};

// Referenced in opts_check_x86.cpp
void ClampX_ClampY_nofilter_scale(const SkBitmapProcState& s, uint32_t xy[],
                                  int count, int x, int y) {
    return NoFilterProc_Scale<ClampTileProcs, true>(s, xy, count, x, y);
}
void ClampX_ClampY_nofilter_affine(const SkBitmapProcState& s, uint32_t xy[],
                                  int count, int x, int y) {
    return NoFilterProc_Affine<ClampTileProcs>(s, xy, count, x, y);
}

static SkBitmapProcState::MatrixProc ClampX_ClampY_Procs[] = {
    // only clamp lives in the right coord space to check for decal
    ClampX_ClampY_nofilter_scale,
    ClampX_ClampY_filter_scale,
    ClampX_ClampY_nofilter_affine,
    ClampX_ClampY_filter_affine,
    NoFilterProc_Persp<ClampTileProcs>,
    ClampX_ClampY_filter_persp
};

#define MAKENAME(suffix)        RepeatX_RepeatY ## suffix
#define TILEX_PROCF(fx, max)    SK_USHIFT16(((fx) & 0xFFFF) * ((max) + 1))
#define TILEY_PROCF(fy, max)    SK_USHIFT16(((fy) & 0xFFFF) * ((max) + 1))
#define TILEX_LOW_BITS(fx, max) ((((fx) & 0xFFFF) * ((max) + 1) >> 12) & 0xF)
#define TILEY_LOW_BITS(fy, max) ((((fy) & 0xFFFF) * ((max) + 1) >> 12) & 0xF)
#include "SkBitmapProcState_matrix.h"

struct RepeatTileProcs {
    static unsigned X(const SkBitmapProcState&, SkFixed fx, int max) {
        return SK_USHIFT16(((fx) & 0xFFFF) * ((max) + 1));
    }
    static unsigned Y(const SkBitmapProcState&, SkFixed fy, int max) {
        return SK_USHIFT16(((fy) & 0xFFFF) * ((max) + 1));
    }
};

static SkBitmapProcState::MatrixProc RepeatX_RepeatY_Procs[] = {
    NoFilterProc_Scale<RepeatTileProcs, false>,
    RepeatX_RepeatY_filter_scale,
    NoFilterProc_Affine<RepeatTileProcs>,
    RepeatX_RepeatY_filter_affine,
    NoFilterProc_Persp<RepeatTileProcs>,
    RepeatX_RepeatY_filter_persp
};
#endif

#define MAKENAME(suffix)        GeneralXY ## suffix
#define PREAMBLE(state)         SkBitmapProcState::FixedTileProc tileProcX = (state).fTileProcX; (void) tileProcX; \
                                SkBitmapProcState::FixedTileProc tileProcY = (state).fTileProcY; (void) tileProcY; \
                                SkBitmapProcState::FixedTileLowBitsProc tileLowBitsProcX = (state).fTileLowBitsProcX; (void) tileLowBitsProcX; \
                                SkBitmapProcState::FixedTileLowBitsProc tileLowBitsProcY = (state).fTileLowBitsProcY; (void) tileLowBitsProcY
#define PREAMBLE_PARAM_X        , SkBitmapProcState::FixedTileProc tileProcX, SkBitmapProcState::FixedTileLowBitsProc tileLowBitsProcX
#define PREAMBLE_PARAM_Y        , SkBitmapProcState::FixedTileProc tileProcY, SkBitmapProcState::FixedTileLowBitsProc tileLowBitsProcY
#define PREAMBLE_ARG_X          , tileProcX, tileLowBitsProcX
#define PREAMBLE_ARG_Y          , tileProcY, tileLowBitsProcY
#define TILEX_PROCF(fx, max)    SK_USHIFT16(tileProcX(fx) * ((max) + 1))
#define TILEY_PROCF(fy, max)    SK_USHIFT16(tileProcY(fy) * ((max) + 1))
#define TILEX_LOW_BITS(fx, max) tileLowBitsProcX(fx, (max) + 1)
#define TILEY_LOW_BITS(fy, max) tileLowBitsProcY(fy, (max) + 1)
#include "SkBitmapProcState_matrix.h"

struct GeneralTileProcs {
    static unsigned X(const SkBitmapProcState& s, SkFixed fx, int max) {
        return SK_USHIFT16(s.fTileProcX(fx) * ((max) + 1));
    }
    static unsigned Y(const SkBitmapProcState& s, SkFixed fy, int max) {
        return SK_USHIFT16(s.fTileProcY(fy) * ((max) + 1));
    }
};

static SkBitmapProcState::MatrixProc GeneralXY_Procs[] = {
    NoFilterProc_Scale<GeneralTileProcs, false>,
    GeneralXY_filter_scale,
    NoFilterProc_Affine<GeneralTileProcs>,
    GeneralXY_filter_affine,
    NoFilterProc_Persp<GeneralTileProcs>,
    GeneralXY_filter_persp
};

///////////////////////////////////////////////////////////////////////////////

static inline U16CPU fixed_clamp(SkFixed x) {
    if (x < 0) {
        x = 0;
    }
    if (x >> 16) {
        x = 0xFFFF;
    }
    return x;
}

static inline U16CPU fixed_repeat(SkFixed x) {
    return x & 0xFFFF;
}

// Visual Studio 2010 (MSC_VER=1600) optimizes bit-shift code incorrectly.
// See http://code.google.com/p/skia/issues/detail?id=472
#if defined(_MSC_VER) && (_MSC_VER >= 1600)
#pragma optimize("", off)
#endif

static inline U16CPU fixed_mirror(SkFixed x) {
    SkFixed s = x << 15 >> 31;
    // s is FFFFFFFF if we're on an odd interval, or 0 if an even interval
    return (x ^ s) & 0xFFFF;
}

#if defined(_MSC_VER) && (_MSC_VER >= 1600)
#pragma optimize("", on)
#endif

static SkBitmapProcState::FixedTileProc choose_tile_proc(unsigned m) {
    if (SkShader::kClamp_TileMode == m) {
        return fixed_clamp;
    }
    if (SkShader::kRepeat_TileMode == m) {
        return fixed_repeat;
    }
    SkASSERT(SkShader::kMirror_TileMode == m);
    return fixed_mirror;
}

static inline U16CPU fixed_clamp_lowbits(SkFixed x, int) {
    return (x >> 12) & 0xF;
}

static inline U16CPU fixed_repeat_or_mirrow_lowbits(SkFixed x, int scale) {
    return ((x * scale) >> 12) & 0xF;
}

static SkBitmapProcState::FixedTileLowBitsProc choose_tile_lowbits_proc(unsigned m) {
    if (SkShader::kClamp_TileMode == m) {
        return fixed_clamp_lowbits;
    } else {
        SkASSERT(SkShader::kMirror_TileMode == m ||
                 SkShader::kRepeat_TileMode == m);
        // mirror and repeat have the same behavior for the low bits.
        return fixed_repeat_or_mirrow_lowbits;
    }
}

static inline U16CPU int_clamp(int x, int n) {
    if (x >= n) {
        x = n - 1;
    }
    if (x < 0) {
        x = 0;
    }
    return x;
}

static inline U16CPU int_repeat(int x, int n) {
    return sk_int_mod(x, n);
}

static inline U16CPU int_mirror(int x, int n) {
    x = sk_int_mod(x, 2 * n);
    if (x >= n) {
        x = n + ~(x - n);
    }
    return x;
}

#if 0
static void test_int_tileprocs() {
    for (int i = -8; i <= 8; i++) {
        SkDebugf(" int_mirror(%2d, 3) = %d\n", i, int_mirror(i, 3));
    }
}
#endif

static SkBitmapProcState::IntTileProc choose_int_tile_proc(unsigned tm) {
    if (SkShader::kClamp_TileMode == tm)
        return int_clamp;
    if (SkShader::kRepeat_TileMode == tm)
        return int_repeat;
    SkASSERT(SkShader::kMirror_TileMode == tm);
    return int_mirror;
}

//////////////////////////////////////////////////////////////////////////////

void decal_nofilter_scale(uint32_t dst[], SkFixed fx, SkFixed dx, int count) {
    int i;

    for (i = (count >> 2); i > 0; --i) {
        *dst++ = pack_two_shorts(fx >> 16, (fx + dx) >> 16);
        fx += dx+dx;
        *dst++ = pack_two_shorts(fx >> 16, (fx + dx) >> 16);
        fx += dx+dx;
    }
    count &= 3;

    uint16_t* xx = (uint16_t*)dst;
    for (i = count; i > 0; --i) {
        *xx++ = SkToU16(fx >> 16); fx += dx;
    }
}

void decal_filter_scale(uint32_t dst[], SkFixed fx, SkFixed dx, int count) {
    if (count & 1) {
        SkASSERT((fx >> (16 + 14)) == 0);
        *dst++ = (fx >> 12 << 14) | ((fx >> 16) + 1);
        fx += dx;
    }
    while ((count -= 2) >= 0) {
        SkASSERT((fx >> (16 + 14)) == 0);
        *dst++ = (fx >> 12 << 14) | ((fx >> 16) + 1);
        fx += dx;

        *dst++ = (fx >> 12 << 14) | ((fx >> 16) + 1);
        fx += dx;
    }
}

///////////////////////////////////////////////////////////////////////////////
// stores the same as SCALE, but is cheaper to compute. Also since there is no
// scale, we don't need/have a FILTER version

static void fill_sequential(uint16_t xptr[], int start, int count) {
#if 1
    if (reinterpret_cast<intptr_t>(xptr) & 0x2) {
        *xptr++ = start++;
        count -= 1;
    }
    if (count > 3) {
        uint32_t* xxptr = reinterpret_cast<uint32_t*>(xptr);
        uint32_t pattern0 = PACK_TWO_SHORTS(start + 0, start + 1);
        uint32_t pattern1 = PACK_TWO_SHORTS(start + 2, start + 3);
        start += count & ~3;
        int qcount = count >> 2;
        do {
            *xxptr++ = pattern0;
            pattern0 += 0x40004;
            *xxptr++ = pattern1;
            pattern1 += 0x40004;
        } while (--qcount != 0);
        xptr = reinterpret_cast<uint16_t*>(xxptr);
        count &= 3;
    }
    while (--count >= 0) {
        *xptr++ = start++;
    }
#else
    for (int i = 0; i < count; i++) {
        *xptr++ = start++;
    }
#endif
}

static int nofilter_trans_preamble(const SkBitmapProcState& s, uint32_t** xy,
                                   int x, int y) {
    SkPoint pt;
    s.fInvProc(s.fInvMatrix, SkIntToScalar(x) + SK_ScalarHalf,
               SkIntToScalar(y) + SK_ScalarHalf, &pt);
    **xy = s.fIntTileProcY(SkScalarToFixed(pt.fY) >> 16,
                           s.fBitmap->height());
    *xy += 1;   // bump the ptr
    // return our starting X position
    return SkScalarToFixed(pt.fX) >> 16;
}

static void clampx_nofilter_trans(const SkBitmapProcState& s,
                                  uint32_t xy[], int count, int x, int y) {
    SkASSERT((s.fInvType & ~SkMatrix::kTranslate_Mask) == 0);

    int xpos = nofilter_trans_preamble(s, &xy, x, y);
    const int width = s.fBitmap->width();
    if (1 == width) {
        // all of the following X values must be 0
        memset(xy, 0, count * sizeof(uint16_t));
        return;
    }

    uint16_t* xptr = reinterpret_cast<uint16_t*>(xy);
    int n;

    // fill before 0 as needed
    if (xpos < 0) {
        n = -xpos;
        if (n > count) {
            n = count;
        }
        memset(xptr, 0, n * sizeof(uint16_t));
        count -= n;
        if (0 == count) {
            return;
        }
        xptr += n;
        xpos = 0;
    }

    // fill in 0..width-1 if needed
    if (xpos < width) {
        n = width - xpos;
        if (n > count) {
            n = count;
        }
        fill_sequential(xptr, xpos, n);
        count -= n;
        if (0 == count) {
            return;
        }
        xptr += n;
    }

    // fill the remaining with the max value
    sk_memset16(xptr, width - 1, count);
}

static void repeatx_nofilter_trans(const SkBitmapProcState& s,
                                   uint32_t xy[], int count, int x, int y) {
    SkASSERT((s.fInvType & ~SkMatrix::kTranslate_Mask) == 0);

    int xpos = nofilter_trans_preamble(s, &xy, x, y);
    const int width = s.fBitmap->width();
    if (1 == width) {
        // all of the following X values must be 0
        memset(xy, 0, count * sizeof(uint16_t));
        return;
    }

    uint16_t* xptr = reinterpret_cast<uint16_t*>(xy);
    int start = sk_int_mod(xpos, width);
    int n = width - start;
    if (n > count) {
        n = count;
    }
    fill_sequential(xptr, start, n);
    xptr += n;
    count -= n;

    while (count >= width) {
        fill_sequential(xptr, 0, width);
        xptr += width;
        count -= width;
    }

    if (count > 0) {
        fill_sequential(xptr, 0, count);
    }
}

static void fill_backwards(uint16_t xptr[], int pos, int count) {
    for (int i = 0; i < count; i++) {
        SkASSERT(pos >= 0);
        xptr[i] = pos--;
    }
}

static void mirrorx_nofilter_trans(const SkBitmapProcState& s,
                                   uint32_t xy[], int count, int x, int y) {
    SkASSERT((s.fInvType & ~SkMatrix::kTranslate_Mask) == 0);

    int xpos = nofilter_trans_preamble(s, &xy, x, y);
    const int width = s.fBitmap->width();
    if (1 == width) {
        // all of the following X values must be 0
        memset(xy, 0, count * sizeof(uint16_t));
        return;
    }

    uint16_t* xptr = reinterpret_cast<uint16_t*>(xy);
    // need to know our start, and our initial phase (forward or backward)
    bool forward;
    int n;
    int start = sk_int_mod(xpos, 2 * width);
    if (start >= width) {
        start = width + ~(start - width);
        forward = false;
        n = start + 1;  // [start .. 0]
    } else {
        forward = true;
        n = width - start;  // [start .. width)
    }
    if (n > count) {
        n = count;
    }
    if (forward) {
        fill_sequential(xptr, start, n);
    } else {
        fill_backwards(xptr, start, n);
    }
    forward = !forward;
    xptr += n;
    count -= n;

    while (count >= width) {
        if (forward) {
            fill_sequential(xptr, 0, width);
        } else {
            fill_backwards(xptr, width - 1, width);
        }
        forward = !forward;
        xptr += width;
        count -= width;
    }

    if (count > 0) {
        if (forward) {
            fill_sequential(xptr, 0, count);
        } else {
            fill_backwards(xptr, width - 1, count);
        }
    }
}

///////////////////////////////////////////////////////////////////////////////

SkBitmapProcState::MatrixProc SkBitmapProcState::chooseMatrixProc(bool trivial_matrix) {
//    test_int_tileprocs();
    // check for our special case when there is no scale/affine/perspective
    if (trivial_matrix) {
        SkASSERT(SkPaint::kNone_FilterLevel == fFilterLevel);
        fIntTileProcY = choose_int_tile_proc(fTileModeY);
        switch (fTileModeX) {
            case SkShader::kClamp_TileMode:
                return clampx_nofilter_trans;
            case SkShader::kRepeat_TileMode:
                return repeatx_nofilter_trans;
            case SkShader::kMirror_TileMode:
                return mirrorx_nofilter_trans;
        }
    }

    int index = 0;
    if (fFilterLevel != SkPaint::kNone_FilterLevel) {
        index = 1;
    }
    if (fInvType & SkMatrix::kPerspective_Mask) {
        index += 4;
    } else if (fInvType & SkMatrix::kAffine_Mask) {
        index += 2;
    }

    if (SkShader::kClamp_TileMode == fTileModeX && SkShader::kClamp_TileMode == fTileModeY) {
        // clamp gets special version of filterOne
        fFilterOneX = SK_Fixed1;
        fFilterOneY = SK_Fixed1;
        return SK_ARM_NEON_WRAP(ClampX_ClampY_Procs)[index];
    }

    // all remaining procs use this form for filterOne
    fFilterOneX = SK_Fixed1 / fBitmap->width();
    fFilterOneY = SK_Fixed1 / fBitmap->height();

    if (SkShader::kRepeat_TileMode == fTileModeX && SkShader::kRepeat_TileMode == fTileModeY) {
        return SK_ARM_NEON_WRAP(RepeatX_RepeatY_Procs)[index];
    }

    fTileProcX = choose_tile_proc(fTileModeX);
    fTileProcY = choose_tile_proc(fTileModeY);
    fTileLowBitsProcX = choose_tile_lowbits_proc(fTileModeX);
    fTileLowBitsProcY = choose_tile_lowbits_proc(fTileModeY);
    return GeneralXY_Procs[index];
}