/*
 * By downloading, copying, installing or using the software you agree to this license.
 * If you do not agree to this license, do not download, install,
 * copy or use the software.
 *
 *
 *                           License Agreement
 *                For Open Source Computer Vision Library
 *                        (3-clause BSD License)
 *
 * Copyright (C) 2012-2015, NVIDIA Corporation, all rights reserved.
 * Third party copyrights are property of their respective owners.
 *
 * Redistribution and use in source and binary forms, with or without modification,
 * are permitted provided that the following conditions are met:
 *
 *   * Redistributions of source code must retain the above copyright notice,
 *     this list of conditions and the following disclaimer.
 *
 *   * Redistributions in binary form must reproduce the above copyright notice,
 *     this list of conditions and the following disclaimer in the documentation
 *     and/or other materials provided with the distribution.
 *
 *   * Neither the names of the copyright holders nor the names of the contributors
 *     may be used to endorse or promote products derived from this software
 *     without specific prior written permission.
 *
 * This software is provided by the copyright holders and contributors "as is" and
 * any express or implied warranties, including, but not limited to, the implied
 * warranties of merchantability and fitness for a particular purpose are disclaimed.
 * In no event shall copyright holders or contributors be liable for any direct,
 * indirect, incidental, special, exemplary, or consequential damages
 * (including, but not limited to, procurement of substitute goods or services;
 * loss of use, data, or profits; or business interruption) however caused
 * and on any theory of liability, whether in contract, strict liability,
 * or tort (including negligence or otherwise) arising in any way out of
 * the use of this software, even if advised of the possibility of such damage.
 */

#include "common.hpp"

#include "vtransform.hpp"

namespace CAROTENE_NS {

#ifdef CAROTENE_NEON

namespace {

inline void vnst(u8* dst, uint8x16_t v1, uint8x16_t v2) { vst1q_u8(dst, v1); vst1q_u8(dst+16, v2); }
inline void vnst(u8* dst, uint16x8_t v1, uint16x8_t v2) { vst1q_u8(dst, vcombine_u8(vmovn_u16(v1), vmovn_u16(v2))); }
inline void vnst(u8* dst, uint32x4_t v1, uint32x4_t v2) { vst1_u8(dst, vmovn_u16(vcombine_u16(vmovn_u32(v1), vmovn_u32(v2)))); }

template <typename T, int elsize> struct vtail
{
    static inline void inRange(const T *, const T *, const T *,
                               u8 *, size_t &, size_t)
    {
        //do nothing since there couldn't be enough data
    }
};
template <typename T> struct vtail<T, 2>
{
    static inline void inRange(const T * src, const T * rng1, const T * rng2,
                               u8 * dst, size_t &x, size_t width)
    {
        typedef typename internal::VecTraits<T>::vec128 vec128;
        typedef typename internal::VecTraits<T>::unsign::vec128 uvec128;
        //There no more than 15 elements in the tail, so we could handle 8 element vector only once
        if( x + 8 < width)
        {
             vec128  vs = internal::vld1q( src + x);
             vec128 vr1 = internal::vld1q(rng1 + x);
             vec128 vr2 = internal::vld1q(rng2 + x);
            uvec128  vd = internal::vandq(internal::vcgeq(vs, vr1), internal::vcgeq(vr2, vs));
            internal::vst1(dst + x, internal::vmovn(vd));
            x+=8;
        }
    }
};
template <typename T> struct vtail<T, 1>
{
    static inline void inRange(const T * src, const T * rng1, const T * rng2,
                               u8 * dst, size_t &x, size_t width)
    {
        typedef typename internal::VecTraits<T>::vec128 vec128;
        typedef typename internal::VecTraits<T>::unsign::vec128 uvec128;
        typedef typename internal::VecTraits<T>::vec64 vec64;
        typedef typename internal::VecTraits<T>::unsign::vec64 uvec64;
        //There no more than 31 elements in the tail, so we could handle once 16+8 or 16 or 8 elements
        if( x + 16 < width)
        {
             vec128  vs = internal::vld1q( src + x);
             vec128 vr1 = internal::vld1q(rng1 + x);
             vec128 vr2 = internal::vld1q(rng2 + x);
            uvec128  vd = internal::vandq(internal::vcgeq(vs, vr1), internal::vcgeq(vr2, vs));
            internal::vst1q(dst + x, vd);
            x+=16;
        }
        if( x + 8 < width)
        {
             vec64  vs = internal::vld1( src + x);
             vec64 vr1 = internal::vld1(rng1 + x);
             vec64 vr2 = internal::vld1(rng2 + x);
            uvec64  vd = internal::vand(internal::vcge(vs, vr1), internal::vcge(vr2, vs));
            internal::vst1(dst + x, vd);
            x+=8;
        }
    }
};

template <typename T>
inline void inRangeCheck(const Size2D &_size,
                         const T * srcBase, ptrdiff_t srcStride,
                         const T * rng1Base, ptrdiff_t rng1Stride,
                         const T * rng2Base, ptrdiff_t rng2Stride,
                         u8 * dstBase, ptrdiff_t dstStride)
{
    typedef typename internal::VecTraits<T>::vec128 vec128;
    typedef typename internal::VecTraits<T>::unsign::vec128 uvec128;

    Size2D size(_size);
    if (srcStride == dstStride &&
        srcStride == rng1Stride &&
        srcStride == rng2Stride &&
        srcStride == (ptrdiff_t)(size.width))
    {
        size.width *= size.height;
        size.height = 1;
    }
    const size_t width = size.width & ~( 32/sizeof(T) - 1 );

    for(size_t j = 0; j < size.height; ++j)
    {
        const T *  src = internal::getRowPtr( srcBase,  srcStride, j);
        const T * rng1 = internal::getRowPtr(rng1Base, rng1Stride, j);
        const T * rng2 = internal::getRowPtr(rng2Base, rng2Stride, j);
             u8 *  dst = internal::getRowPtr( dstBase,  dstStride, j);
        size_t i = 0;
        for( ; i < width; i += 32/sizeof(T) )
        {
            internal::prefetch(src + i);
            internal::prefetch(rng1 + i);
            internal::prefetch(rng2 + i);

             vec128  vs = internal::vld1q( src + i);
             vec128 vr1 = internal::vld1q(rng1 + i);
             vec128 vr2 = internal::vld1q(rng2 + i);
            uvec128 vd1 = internal::vandq(internal::vcgeq(vs, vr1), internal::vcgeq(vr2, vs));
                     vs = internal::vld1q( src + i + 16/sizeof(T));
                    vr1 = internal::vld1q(rng1 + i + 16/sizeof(T));
                    vr2 = internal::vld1q(rng2 + i + 16/sizeof(T));
            uvec128 vd2 = internal::vandq(internal::vcgeq(vs, vr1), internal::vcgeq(vr2, vs));
            vnst(dst + i, vd1, vd2);
        }
        vtail<T, sizeof(T)>::inRange(src, rng1, rng2, dst, i, size.width);
        for( ; i < size.width; i++ )
            dst[i] = (u8)(-(rng1[i] <= src[i] && src[i] <= rng2[i]));
    }
}

}

#define INRANGEFUNC(T)                                       \
void inRange(const Size2D &_size,                            \
             const T * srcBase, ptrdiff_t srcStride,         \
             const T * rng1Base, ptrdiff_t rng1Stride,       \
             const T * rng2Base, ptrdiff_t rng2Stride,       \
             u8 * dstBase, ptrdiff_t dstStride)              \
{                                                            \
    internal::assertSupportedConfiguration();                \
    inRangeCheck(_size, srcBase, srcStride,                  \
                 rng1Base, rng1Stride, rng2Base, rng2Stride, \
                 dstBase, dstStride);                        \
}
#else
#define INRANGEFUNC(T)                                       \
void inRange(const Size2D &,                                 \
             const T *, ptrdiff_t,                           \
             const T *, ptrdiff_t,                           \
             const T *, ptrdiff_t,                           \
             u8 *, ptrdiff_t)                                \
{                                                            \
    internal::assertSupportedConfiguration();                \
}
#endif

INRANGEFUNC(u8)
INRANGEFUNC(s8)
INRANGEFUNC(u16)
INRANGEFUNC(s16)
INRANGEFUNC(s32)
INRANGEFUNC(f32)

} // namespace CAROTENE_NS