| 1 | /**************************************************************************** | 
|---|
| 2 | ** | 
|---|
| 3 | ** Copyright (C) 2016 The Qt Company Ltd. | 
|---|
| 4 | ** Contact: https://www.qt.io/licensing/ | 
|---|
| 5 | ** | 
|---|
| 6 | ** This file is part of the QtGui module of the Qt Toolkit. | 
|---|
| 7 | ** | 
|---|
| 8 | ** $QT_BEGIN_LICENSE:LGPL$ | 
|---|
| 9 | ** Commercial License Usage | 
|---|
| 10 | ** Licensees holding valid commercial Qt licenses may use this file in | 
|---|
| 11 | ** accordance with the commercial license agreement provided with the | 
|---|
| 12 | ** Software or, alternatively, in accordance with the terms contained in | 
|---|
| 13 | ** a written agreement between you and The Qt Company. For licensing terms | 
|---|
| 14 | ** and conditions see https://www.qt.io/terms-conditions. For further | 
|---|
| 15 | ** information use the contact form at https://www.qt.io/contact-us. | 
|---|
| 16 | ** | 
|---|
| 17 | ** GNU Lesser General Public License Usage | 
|---|
| 18 | ** Alternatively, this file may be used under the terms of the GNU Lesser | 
|---|
| 19 | ** General Public License version 3 as published by the Free Software | 
|---|
| 20 | ** Foundation and appearing in the file LICENSE.LGPL3 included in the | 
|---|
| 21 | ** packaging of this file. Please review the following information to | 
|---|
| 22 | ** ensure the GNU Lesser General Public License version 3 requirements | 
|---|
| 23 | ** will be met: https://www.gnu.org/licenses/lgpl-3.0.html. | 
|---|
| 24 | ** | 
|---|
| 25 | ** GNU General Public License Usage | 
|---|
| 26 | ** Alternatively, this file may be used under the terms of the GNU | 
|---|
| 27 | ** General Public License version 2.0 or (at your option) the GNU General | 
|---|
| 28 | ** Public license version 3 or any later version approved by the KDE Free | 
|---|
| 29 | ** Qt Foundation. The licenses are as published by the Free Software | 
|---|
| 30 | ** Foundation and appearing in the file LICENSE.GPL2 and LICENSE.GPL3 | 
|---|
| 31 | ** included in the packaging of this file. Please review the following | 
|---|
| 32 | ** information to ensure the GNU General Public License requirements will | 
|---|
| 33 | ** be met: https://www.gnu.org/licenses/gpl-2.0.html and | 
|---|
| 34 | ** https://www.gnu.org/licenses/gpl-3.0.html. | 
|---|
| 35 | ** | 
|---|
| 36 | ** $QT_END_LICENSE$ | 
|---|
| 37 | ** | 
|---|
| 38 | ****************************************************************************/ | 
|---|
| 39 |  | 
|---|
| 40 | #include <qimage.h> | 
|---|
| 41 | #include <private/qimage_p.h> | 
|---|
| 42 | #include <private/qsimd_p.h> | 
|---|
| 43 |  | 
|---|
| 44 | #ifdef QT_COMPILER_SUPPORTS_SSSE3 | 
|---|
| 45 |  | 
|---|
| 46 | QT_BEGIN_NAMESPACE | 
|---|
| 47 |  | 
|---|
| 48 | // Convert a scanline of RGB888 (src) to RGB32 (dst) | 
|---|
| 49 | // src must be at least len * 3 bytes | 
|---|
| 50 | // dst must be at least len * 4 bytes | 
|---|
| 51 | Q_GUI_EXPORT void QT_FASTCALL qt_convert_rgb888_to_rgb32_ssse3(quint32 *dst, const uchar *src, int len) | 
|---|
| 52 | { | 
|---|
| 53 | int i = 0; | 
|---|
| 54 |  | 
|---|
| 55 | // Prologue, align dst to 16 bytes. | 
|---|
| 56 | ALIGNMENT_PROLOGUE_16BYTES(dst, i, len) { | 
|---|
| 57 | dst[i] = qRgb(src[0], src[1], src[2]); | 
|---|
| 58 | src += 3; | 
|---|
| 59 | } | 
|---|
| 60 |  | 
|---|
| 61 | // Mask the 4 first colors of the RGB888 vector | 
|---|
| 62 | const __m128i shuffleMask = _mm_set_epi8(char(0xff), 9, 10, 11, char(0xff), 6, 7, 8, char(0xff), 3, 4, 5, char(0xff), 0, 1, 2); | 
|---|
| 63 |  | 
|---|
| 64 | // Mask the 4 last colors of a RGB888 vector with an offset of 1 (so the last 3 bytes are RGB) | 
|---|
| 65 | const __m128i shuffleMaskEnd = _mm_set_epi8(char(0xff), 13, 14, 15, char(0xff), 10, 11, 12, char(0xff), 7, 8, 9, char(0xff), 4, 5, 6); | 
|---|
| 66 |  | 
|---|
| 67 | // Mask to have alpha = 0xff | 
|---|
| 68 | const __m128i alphaMask = _mm_set1_epi32(0xff000000); | 
|---|
| 69 |  | 
|---|
| 70 | const __m128i *inVectorPtr = (const __m128i *)src; | 
|---|
| 71 | __m128i *dstVectorPtr = (__m128i *)(dst + i); | 
|---|
| 72 |  | 
|---|
| 73 | for (; i < (len - 15); i += 16) { // one iteration in the loop converts 16 pixels | 
|---|
| 74 | /* | 
|---|
| 75 | RGB888 has 5 pixels per vector, + 1 byte from the next pixel. The idea here is | 
|---|
| 76 | to load vectors of RGB888 and use palignr to select a vector out of two vectors. | 
|---|
| 77 |  | 
|---|
| 78 | After 3 loads of RGB888 and 3 stores of RGB32, we have 4 pixels left in the last | 
|---|
| 79 | vector of RGB888, we can mask it directly to get a last store or RGB32. After that, | 
|---|
| 80 | the first next byte is a R, and we can loop for the next 16 pixels. | 
|---|
| 81 |  | 
|---|
| 82 | The conversion itself is done with a byte permutation (pshufb). | 
|---|
| 83 | */ | 
|---|
| 84 | __m128i firstSrcVector = _mm_lddqu_si128(inVectorPtr); | 
|---|
| 85 | __m128i outputVector = _mm_shuffle_epi8(firstSrcVector, shuffleMask); | 
|---|
| 86 | _mm_store_si128(dstVectorPtr, _mm_or_si128(outputVector, alphaMask)); | 
|---|
| 87 | ++inVectorPtr; | 
|---|
| 88 | ++dstVectorPtr; | 
|---|
| 89 |  | 
|---|
| 90 | // There are 4 unused bytes left in srcVector, we need to load the next 16 bytes | 
|---|
| 91 | // and load the next input with palignr | 
|---|
| 92 | __m128i secondSrcVector = _mm_lddqu_si128(inVectorPtr); | 
|---|
| 93 | __m128i srcVector = _mm_alignr_epi8(secondSrcVector, firstSrcVector, 12); | 
|---|
| 94 | outputVector = _mm_shuffle_epi8(srcVector, shuffleMask); | 
|---|
| 95 | _mm_store_si128(dstVectorPtr, _mm_or_si128(outputVector, alphaMask)); | 
|---|
| 96 | ++inVectorPtr; | 
|---|
| 97 | ++dstVectorPtr; | 
|---|
| 98 | firstSrcVector = secondSrcVector; | 
|---|
| 99 |  | 
|---|
| 100 | // We now have 8 unused bytes left in firstSrcVector | 
|---|
| 101 | secondSrcVector = _mm_lddqu_si128(inVectorPtr); | 
|---|
| 102 | srcVector = _mm_alignr_epi8(secondSrcVector, firstSrcVector, 8); | 
|---|
| 103 | outputVector = _mm_shuffle_epi8(srcVector, shuffleMask); | 
|---|
| 104 | _mm_store_si128(dstVectorPtr, _mm_or_si128(outputVector, alphaMask)); | 
|---|
| 105 | ++inVectorPtr; | 
|---|
| 106 | ++dstVectorPtr; | 
|---|
| 107 |  | 
|---|
| 108 | // There are now 12 unused bytes in firstSrcVector. | 
|---|
| 109 | // We can mask them directly, almost there. | 
|---|
| 110 | outputVector = _mm_shuffle_epi8(secondSrcVector, shuffleMaskEnd); | 
|---|
| 111 | _mm_store_si128(dstVectorPtr, _mm_or_si128(outputVector, alphaMask)); | 
|---|
| 112 | ++dstVectorPtr; | 
|---|
| 113 | } | 
|---|
| 114 | src = (const uchar *)inVectorPtr; | 
|---|
| 115 |  | 
|---|
| 116 | SIMD_EPILOGUE(i, len, 15) { | 
|---|
| 117 | dst[i] = qRgb(src[0], src[1], src[2]); | 
|---|
| 118 | src += 3; | 
|---|
| 119 | } | 
|---|
| 120 | } | 
|---|
| 121 |  | 
|---|
| 122 | void convert_RGB888_to_RGB32_ssse3(QImageData *dest, const QImageData *src, Qt::ImageConversionFlags) | 
|---|
| 123 | { | 
|---|
| 124 | Q_ASSERT(src->format == QImage::Format_RGB888 || src->format == QImage::Format_BGR888); | 
|---|
| 125 | if (src->format == QImage::Format_BGR888) | 
|---|
| 126 | Q_ASSERT(dest->format == QImage::Format_RGBX8888 || dest->format == QImage::Format_RGBA8888 || dest->format == QImage::Format_RGBA8888_Premultiplied); | 
|---|
| 127 | else | 
|---|
| 128 | Q_ASSERT(dest->format == QImage::Format_RGB32 || dest->format == QImage::Format_ARGB32 || dest->format == QImage::Format_ARGB32_Premultiplied); | 
|---|
| 129 | Q_ASSERT(src->width == dest->width); | 
|---|
| 130 | Q_ASSERT(src->height == dest->height); | 
|---|
| 131 |  | 
|---|
| 132 | const uchar *src_data = (uchar *) src->data; | 
|---|
| 133 | quint32 *dest_data = (quint32 *) dest->data; | 
|---|
| 134 |  | 
|---|
| 135 | for (int i = 0; i < src->height; ++i) { | 
|---|
| 136 | qt_convert_rgb888_to_rgb32_ssse3(dest_data, src_data, src->width); | 
|---|
| 137 | src_data += src->bytes_per_line; | 
|---|
| 138 | dest_data = (quint32 *)((uchar*)dest_data + dest->bytes_per_line); | 
|---|
| 139 | } | 
|---|
| 140 | } | 
|---|
| 141 |  | 
|---|
| 142 | QT_END_NAMESPACE | 
|---|
| 143 |  | 
|---|
| 144 | #endif // QT_COMPILER_SUPPORTS_SSSE3 | 
|---|
| 145 |  | 
|---|