src/core/SkBlitRow_D32.cpp

/*
 * Copyright 2011 Google Inc.
 *
 * Use of this source code is governed by a BSD-style license that can be
 * found in the LICENSE file.
 */

#include "SkBlitRow.h"
#include "SkBlitMask.h"
#include "SkColorPriv.h"
#include "SkUtils.h"

#define UNROLL

static void S32_Opaque_BlitRow32(SkPMColor* SK_RESTRICT dst,
                                 const SkPMColor* SK_RESTRICT src,
                                 int count, U8CPU alpha) {
    SkASSERT(255 == alpha);
    sk_memcpy32(dst, src, count);
}

static void S32_Blend_BlitRow32(SkPMColor* SK_RESTRICT dst,
                                const SkPMColor* SK_RESTRICT src,
                                int count, U8CPU alpha) {
    SkASSERT(alpha <= 255);
    if (count > 0) {
        unsigned src_scale = SkAlpha255To256(alpha);
        unsigned dst_scale = 256 - src_scale;

#ifdef UNROLL
        if (count & 1) {
            *dst = SkAlphaMulQ(*(src++), src_scale) + SkAlphaMulQ(*dst, dst_scale);
            dst += 1;
            count -= 1;
        }

        const SkPMColor* SK_RESTRICT srcEnd = src + count;
        while (src != srcEnd) {
            *dst = SkAlphaMulQ(*(src++), src_scale) + SkAlphaMulQ(*dst, dst_scale);
            dst += 1;
            *dst = SkAlphaMulQ(*(src++), src_scale) + SkAlphaMulQ(*dst, dst_scale);
            dst += 1;
        }
#else
        do {
            *dst = SkAlphaMulQ(*src, src_scale) + SkAlphaMulQ(*dst, dst_scale);
            src += 1;
            dst += 1;
        } while (--count > 0);
#endif
    }
}

static void S32A_Opaque_BlitRow32(SkPMColor* SK_RESTRICT dst,
                                  const SkPMColor* SK_RESTRICT src,
                                  int count, U8CPU alpha) {
    SkASSERT(255 == alpha);
    if (count > 0) {
#ifdef UNROLL
        if (count & 1) {
            *dst = SkPMSrcOver(*(src++), *dst);
            dst += 1;
            count -= 1;
        }

        const SkPMColor* SK_RESTRICT srcEnd = src + count;
        while (src != srcEnd) {
            *dst = SkPMSrcOver(*(src++), *dst);
            dst += 1;
            *dst = SkPMSrcOver(*(src++), *dst);
            dst += 1;
        }
#else
        do {
            *dst = SkPMSrcOver(*src, *dst);
            src += 1;
            dst += 1;
        } while (--count > 0);
#endif
    }
}

static void S32A_Blend_BlitRow32(SkPMColor* SK_RESTRICT dst,
                                 const SkPMColor* SK_RESTRICT src,
                                 int count, U8CPU alpha) {
    SkASSERT(alpha <= 255);
    if (count > 0) {
#ifdef UNROLL
        if (count & 1) {
            *dst = SkBlendARGB32(*(src++), *dst, alpha);
            dst += 1;
            count -= 1;
        }

        const SkPMColor* SK_RESTRICT srcEnd = src + count;
        while (src != srcEnd) {
            *dst = SkBlendARGB32(*(src++), *dst, alpha);
            dst += 1;
            *dst = SkBlendARGB32(*(src++), *dst, alpha);
            dst += 1;
        }
#else
        do {
            *dst = SkBlendARGB32(*src, *dst, alpha);
            src += 1;
            dst += 1;
        } while (--count > 0);
#endif
    }
}

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

static const SkBlitRow::Proc32 gDefault_Procs32[] = {
    S32_Opaque_BlitRow32,
    S32_Blend_BlitRow32,
    S32A_Opaque_BlitRow32,
    S32A_Blend_BlitRow32
};

SkBlitRow::Proc32 SkBlitRow::Factory32(unsigned flags) {
    SkASSERT(flags < SK_ARRAY_COUNT(gDefault_Procs32));
    // just so we don't crash
    flags &= kFlags32_Mask;

    SkBlitRow::Proc32 proc = PlatformProcs32(flags);
    if (NULL == proc) {
        proc = gDefault_Procs32[flags];
    }
    SkASSERT(proc);
    return proc;
}

SkBlitRow::Proc32 SkBlitRow::ColorProcFactory() {
    SkBlitRow::ColorProc proc = PlatformColorProc();
    if (NULL == proc) {
        proc = Color32;
    }
    SkASSERT(proc);
    return proc;
}

#define SK_SUPPORT_LEGACY_COLOR32_MATHx

// Color32 and its SIMD specializations use the blend_256_round_alt algorithm
// from tests/BlendTest.cpp.  It's not quite perfect, but it's never wrong in the
// interesting edge cases, and it's quite a bit faster than blend_perfect.
//
// blend_256_round_alt is our currently blessed algorithm.  Please use it or an analogous one.
void SkBlitRow::Color32(SkPMColor* SK_RESTRICT dst,
                        const SkPMColor* SK_RESTRICT src,
                        int count, SkPMColor color) {
    switch (SkGetPackedA32(color)) {
        case   0: memmove(dst, src, count * sizeof(SkPMColor)); return;
        case 255: sk_memset32(dst, color, count);               return;
    }

    unsigned invA = 255 - SkGetPackedA32(color);
#ifdef SK_SUPPORT_LEGACY_COLOR32_MATH  // blend_256_plus1_trunc, busted
    unsigned round = 0;
#else                          // blend_256_round_alt, good
    invA += invA >> 7;
    unsigned round = (128 << 16) + (128 << 0);
#endif

    while (count --> 0) {
        // Our math is 16-bit, so we can do a little bit of SIMD in 32-bit registers.
        const uint32_t mask = 0x00FF00FF;
        uint32_t rb = (((*src >> 0) & mask) * invA + round) >> 8,  // _r_b
                 ag = (((*src >> 8) & mask) * invA + round) >> 0;  // a_g_
        *dst = color + ((rb & mask) | (ag & ~mask));
        src++;
        dst++;
    }
}