0N/A/*
2362N/A * Copyright (c) 2003, 2008, Oracle and/or its affiliates. All rights reserved.
0N/A * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
0N/A *
0N/A * This code is free software; you can redistribute it and/or modify it
0N/A * under the terms of the GNU General Public License version 2 only, as
2362N/A * published by the Free Software Foundation. Oracle designates this
0N/A * particular file as subject to the "Classpath" exception as provided
2362N/A * by Oracle in the LICENSE file that accompanied this code.
0N/A *
0N/A * This code is distributed in the hope that it will be useful, but WITHOUT
0N/A * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
0N/A * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
0N/A * version 2 for more details (a copy is included in the LICENSE file that
0N/A * accompanied this code).
0N/A *
0N/A * You should have received a copy of the GNU General Public License version
0N/A * 2 along with this work; if not, write to the Free Software Foundation,
0N/A * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
0N/A *
2362N/A * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
2362N/A * or visit www.oracle.com if you need additional information or have any
2362N/A * questions.
0N/A */
0N/A
0N/A#if !defined(JAVA2D_NO_MLIB) || defined(MLIB_ADD_SUFF)
0N/A
0N/A#include "vis_AlphaMacros.h"
0N/A
0N/A/***************************************************************/
0N/A
0N/A#define GET_ARGBPRE(i) \
0N/A 0xFF000000 | (src[3*i + 2] << 16) | (src[3*i + 1] << 8) | src[3*i]
0N/A
0N/A/***************************************************************/
0N/A
0N/A#define CONVERT_PRE(rr, dstA, dstARGB) \
0N/A rr = vis_fmul8x16(dstARGB, ((mlib_d64*)vis_div8pre_tbl)[dstA])
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntArgbPreToIntArgbConvert)(BLIT_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 dstA0, dstA1;
0N/A mlib_d64 res0, res1, dstARGB;
0N/A mlib_f32 dstARGB0;
0N/A mlib_s32 i, i0, j;
0N/A
0N/A vis_write_gsr(7 << 3);
0N/A
0N/A if (dstScan == 4*width && srcScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstARGB0 = src[i];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstA1 = *(mlib_u8*)(src + i + 1);
0N/A dstARGB = vis_freg_pair(src[i], src[i + 1]);
0N/A
0N/A CONVERT_PRE(res0, dstA0, vis_read_hi(dstARGB));
0N/A CONVERT_PRE(res1, dstA1, vis_read_lo(dstARGB));
0N/A
0N/A res0 = vis_fpack16_pair(res0, res1);
0N/A
0N/A *(mlib_d64*)(dst + i) = res0;
0N/A }
0N/A
0N/A if (i < width) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstARGB0 = src[i];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntArgbPreToIntArgbScaleConvert)(SCALE_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 dstA0, dstA1;
0N/A mlib_d64 res0, res1, dstARGB;
0N/A mlib_f32 dstARGB0;
0N/A mlib_s32 i, i0, j, ind0, ind1;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_s32 *src = srcBase;
0N/A mlib_u8 *dst = dstBase;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A mlib_u32 argb = src[tmpsxloc >> shift];
0N/A mlib_u32 a, r, g, b;
0N/A b = argb & 0xff;
0N/A g = (argb >> 8) & 0xff;
0N/A r = (argb >> 16) & 0xff;
0N/A a = argb >> 24;
0N/A dst[4*i] = a;
0N/A if (a == 0) a = 255; /* a |= (a - 1) >> 24; */
0N/A dst[4*i + 1] = div8table[a][r];
0N/A dst[4*i + 2] = div8table[a][g];
0N/A dst[4*i + 3] = div8table[a][b];
0N/A tmpsxloc += sxinc;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A return;
0N/A }
0N/A
0N/A vis_write_gsr(7 << 3);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A ind0 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A dstA0 = *(mlib_u8*)(src + ind0);
0N/A dstARGB0 = src[ind0];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A ind0 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A ind1 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A dstA0 = *(mlib_u8*)(src + ind0);
0N/A dstA1 = *(mlib_u8*)(src + ind1);
0N/A
0N/A dstARGB = vis_freg_pair(src[ind0], src[ind1]);
0N/A
0N/A CONVERT_PRE(res0, dstA0, vis_read_hi(dstARGB));
0N/A CONVERT_PRE(res1, dstA1, vis_read_lo(dstARGB));
0N/A
0N/A res0 = vis_fpack16_pair(res0, res1);
0N/A
0N/A *(mlib_d64*)(dst + i) = res0;
0N/A }
0N/A
0N/A if (i < width) {
0N/A ind0 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A dstA0 = *(mlib_u8*)(src + ind0);
0N/A dstARGB0 = src[ind0];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/A#undef CONVERT_PRE
0N/A#define CONVERT_PRE(rr, dstA, dstARGB) \
0N/A rr = MUL8_VIS(dstARGB, dstA)
0N/A
0N/Avoid ADD_SUFF(IntArgbToIntArgbPreConvert)(BLIT_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 dstA0, dstA1;
0N/A mlib_d64 res0, res1, dstARGB;
0N/A mlib_f32 dstARGB0;
0N/A mlib_s32 i, i0, j;
0N/A
0N/A vis_write_gsr(0 << 3);
0N/A
0N/A if (dstScan == 4*width && srcScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstARGB0 = src[i];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A *(mlib_u8*)(dst + i) = dstA0;
0N/A
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstA1 = *(mlib_u8*)(src + i + 1);
0N/A dstARGB = vis_freg_pair(src[i], src[i + 1]);
0N/A
0N/A CONVERT_PRE(res0, dstA0, vis_read_hi(dstARGB));
0N/A CONVERT_PRE(res1, dstA1, vis_read_lo(dstARGB));
0N/A
0N/A res0 = vis_fpack16_pair(res0, res1);
0N/A
0N/A *(mlib_d64*)(dst + i) = res0;
0N/A vis_pst_8(dstARGB, dst + i, 0x88);
0N/A }
0N/A
0N/A if (i < width) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstARGB0 = src[i];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A *(mlib_u8*)(dst + i) = dstA0;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntArgbToIntArgbPreScaleConvert)(SCALE_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 dstA0, dstA1;
0N/A mlib_d64 res0, res1, dstARGB;
0N/A mlib_f32 dstARGB0;
0N/A mlib_s32 i, i0, j, ind0, ind1;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_s32 *src = srcBase;
0N/A mlib_u8 *dst = dstBase;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A mlib_u32 argb = src[tmpsxloc >> shift];
0N/A mlib_u32 a, r, g, b;
0N/A b = argb & 0xff;
0N/A g = (argb >> 8) & 0xff;
0N/A r = (argb >> 16) & 0xff;
0N/A a = argb >> 24;
0N/A dst[4*i] = a;
0N/A dst[4*i + 1] = mul8table[a][r];
0N/A dst[4*i + 2] = mul8table[a][g];
0N/A dst[4*i + 3] = mul8table[a][b];
0N/A tmpsxloc += sxinc;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A return;
0N/A }
0N/A
0N/A vis_write_gsr(0 << 3);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A ind0 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A dstA0 = *(mlib_u8*)(src + ind0);
0N/A dstARGB0 = src[ind0];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A *(mlib_u8*)(dst + i) = dstA0;
0N/A
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A ind0 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A ind1 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A dstA0 = *(mlib_u8*)(src + ind0);
0N/A dstA1 = *(mlib_u8*)(src + ind1);
0N/A
0N/A dstARGB = vis_freg_pair(src[ind0], src[ind1]);
0N/A
0N/A CONVERT_PRE(res0, dstA0, vis_read_hi(dstARGB));
0N/A CONVERT_PRE(res1, dstA1, vis_read_lo(dstARGB));
0N/A
0N/A res0 = vis_fpack16_pair(res0, res1);
0N/A
0N/A *(mlib_d64*)(dst + i) = res0;
0N/A vis_pst_8(dstARGB, dst + i, 0x88);
0N/A }
0N/A
0N/A if (i < width) {
0N/A ind0 = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A dstA0 = *(mlib_u8*)(src + ind0);
0N/A dstARGB0 = src[ind0];
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A dst[i] = vis_fpack16(res0);
0N/A *(mlib_u8*)(dst + i) = dstA0;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntArgbToIntArgbPreXorBlit)(BLIT_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 xorpixel = pCompInfo->details.xorPixel;
0N/A mlib_s32 alphamask = pCompInfo->alphaMask;
0N/A mlib_s32 dstA0, dstA1;
0N/A mlib_d64 res0, res1, dstARGB, dd, d_xorpixel, d_alphamask, maskRGB;
0N/A mlib_d64 d_round;
0N/A mlib_f32 dstARGB0, ff;
0N/A mlib_s32 i, i0, j;
0N/A
0N/A vis_write_gsr(0 << 3);
0N/A
0N/A if (dstScan == 4*width && srcScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A d_xorpixel = vis_to_double_dup(xorpixel);
0N/A d_alphamask = vis_to_double_dup(alphamask);
0N/A maskRGB = vis_to_double_dup(0xFFFFFF);
0N/A d_round = vis_to_double_dup(((1 << 16) | 1) << 6);
0N/A
0N/A xorpixel >>= 24;
0N/A alphamask >>= 24;
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstARGB0 = src[i];
0N/A if (dstA0 & 0x80) {
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A res0 = vis_fpadd16(res0, d_round);
0N/A ff = vis_fpack16(res0);
0N/A ff = vis_fxors(ff, vis_read_hi(d_xorpixel));
0N/A ff = vis_fandnots(vis_read_hi(d_alphamask), ff);
0N/A ff = vis_fxors(ff, dst[i]);
0N/A dstA0 = *(mlib_u8*)(dst + i) ^
0N/A ((dstA0 ^ xorpixel) &~ alphamask);
0N/A dst[i] = ff;
0N/A *(mlib_u8*)(dst + i) = dstA0;
0N/A }
0N/A
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstA1 = *(mlib_u8*)(src + i + 1);
0N/A dstARGB = vis_freg_pair(src[i], src[i + 1]);
0N/A
0N/A CONVERT_PRE(res0, dstA0, vis_read_hi(dstARGB));
0N/A CONVERT_PRE(res1, dstA1, vis_read_lo(dstARGB));
0N/A res0 = vis_fpadd16(res0, d_round);
0N/A res1 = vis_fpadd16(res1, d_round);
0N/A dd = vis_fpack16_pair(res0, res1);
0N/A
0N/A dd = vis_for(vis_fand(maskRGB, dd), vis_fandnot(maskRGB, dstARGB));
0N/A
0N/A dd = vis_fxor(dd, d_xorpixel);
0N/A dd = vis_fandnot(d_alphamask, dd);
0N/A dd = vis_fxor(dd, *(mlib_d64*)(dst + i));
0N/A
0N/A vis_pst_32(dd, dst + i, ((dstA0 >> 6) & 2) | (dstA1 >> 7));
0N/A }
0N/A
0N/A if (i < width) {
0N/A dstA0 = *(mlib_u8*)(src + i);
0N/A dstARGB0 = src[i];
0N/A if (dstA0 & 0x80) {
0N/A CONVERT_PRE(res0, dstA0, dstARGB0);
0N/A res0 = vis_fpadd16(res0, d_round);
0N/A ff = vis_fpack16(res0);
0N/A ff = vis_fxors(ff, vis_read_hi(d_xorpixel));
0N/A ff = vis_fandnots(vis_read_hi(d_alphamask), ff);
0N/A ff = vis_fxors(ff, dst[i]);
0N/A dstA0 = *(mlib_u8*)(dst + i) ^
0N/A ((dstA0 ^ xorpixel) &~ alphamask);
0N/A dst[i] = ff;
0N/A *(mlib_u8*)(dst + i) = dstA0;
0N/A }
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntRgbToIntArgbPreConvert)(BLIT_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 dd, mask;
0N/A mlib_s32 i, i0, j;
0N/A
0N/A if (dstScan == 4*width && srcScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A mask = vis_to_double_dup(0xFF000000);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A dst[i] = vis_fors(src[i], vis_read_hi(mask));
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A dd = vis_freg_pair(src[i], src[i + 1]);
0N/A
0N/A *(mlib_d64*)(dst + i) = vis_for(dd, mask);
0N/A }
0N/A
0N/A if (i < width) {
0N/A dst[i] = vis_fors(src[i], vis_read_hi(mask));
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntRgbToIntArgbPreScaleConvert)(SCALE_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 dd, mask;
0N/A mlib_s32 j;
0N/A
0N/A mask = vis_to_double_dup(0xFF000000);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_f32 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A mlib_f32 *dst_end = dst + width;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A *dst++ = vis_fors(src[tmpsxloc >> shift], vis_read_hi(mask));
0N/A tmpsxloc += sxinc;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (; dst <= dst_end - 2; dst += 2) {
0N/A dd = vis_freg_pair(src[tmpsxloc >> shift],
0N/A src[(tmpsxloc + sxinc) >> shift]);
0N/A *(mlib_d64*)dst = vis_for(dd, mask);
0N/A tmpsxloc += 2*sxinc;
0N/A }
0N/A
0N/A if (dst < dst_end) {
0N/A *dst = vis_fors(src[tmpsxloc >> shift], vis_read_hi(mask));
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/A#define BGR_TO_ARGB { \
0N/A mlib_d64 sda, sdb, sdc, sdd, sde, sdf; \
0N/A mlib_d64 s_1, s_2, s_3, a13, b13, a02, b02; \
0N/A \
0N/A sda = vis_fpmerge(vis_read_hi(sd0), vis_read_lo(sd1)); \
0N/A sdb = vis_fpmerge(vis_read_lo(sd0), vis_read_hi(sd2)); \
0N/A sdc = vis_fpmerge(vis_read_hi(sd1), vis_read_lo(sd2)); \
0N/A \
0N/A sdd = vis_fpmerge(vis_read_hi(sda), vis_read_lo(sdb)); \
0N/A sde = vis_fpmerge(vis_read_lo(sda), vis_read_hi(sdc)); \
0N/A sdf = vis_fpmerge(vis_read_hi(sdb), vis_read_lo(sdc)); \
0N/A \
0N/A s_3 = vis_fpmerge(vis_read_hi(sdd), vis_read_lo(sde)); \
0N/A s_2 = vis_fpmerge(vis_read_lo(sdd), vis_read_hi(sdf)); \
0N/A s_1 = vis_fpmerge(vis_read_hi(sde), vis_read_lo(sdf)); \
0N/A \
0N/A a13 = vis_fpmerge(vis_read_hi(s_1), vis_read_hi(s_3)); \
0N/A b13 = vis_fpmerge(vis_read_lo(s_1), vis_read_lo(s_3)); \
0N/A a02 = vis_fpmerge(vis_read_hi(s_0), vis_read_hi(s_2)); \
0N/A b02 = vis_fpmerge(vis_read_lo(s_0), vis_read_lo(s_2)); \
0N/A \
0N/A dd0 = vis_fpmerge(vis_read_hi(a02), vis_read_hi(a13)); \
0N/A dd1 = vis_fpmerge(vis_read_lo(a02), vis_read_lo(a13)); \
0N/A dd2 = vis_fpmerge(vis_read_hi(b02), vis_read_hi(b13)); \
0N/A dd3 = vis_fpmerge(vis_read_lo(b02), vis_read_lo(b13)); \
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ThreeByteBgrToIntArgbPreConvert)(BLIT_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 *sp;
0N/A mlib_d64 s_0;
0N/A mlib_d64 s0, s1, s2, s3, sd0, sd1, sd2, dd0, dd1, dd2, dd3;
0N/A mlib_s32 i, i0, j;
0N/A
0N/A if (srcScan == 3*width && dstScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A s_0 = vis_fone();
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A ((mlib_s32*)dst)[i] = GET_ARGBPRE(i);
0N/A i0 = 1;
0N/A }
0N/A
0N/A sp = vis_alignaddr(src, 3*i0);
0N/A s3 = *sp++;
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 8; i += 8) {
0N/A s0 = s3;
0N/A s1 = *sp++;
0N/A s2 = *sp++;
0N/A s3 = *sp++;
0N/A sd0 = vis_faligndata(s0, s1);
0N/A sd1 = vis_faligndata(s1, s2);
0N/A sd2 = vis_faligndata(s2, s3);
0N/A
0N/A BGR_TO_ARGB
0N/A
0N/A *(mlib_d64*)(dst + i ) = dd0;
0N/A *(mlib_d64*)(dst + i + 2) = dd1;
0N/A *(mlib_d64*)(dst + i + 4) = dd2;
0N/A *(mlib_d64*)(dst + i + 6) = dd3;
0N/A }
0N/A
0N/A for (; i < width; i++) {
0N/A ((mlib_s32*)dst)[i] = GET_ARGBPRE(i);
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ThreeByteBgrToIntArgbPreScaleConvert)(SCALE_PARAMS)
0N/A{
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 dd, maskFF;
0N/A mlib_s32 i, i0, i1, j;
0N/A
0N/A maskFF = vis_fone();
0N/A
0N/A vis_alignaddr(NULL, 7);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_f32 *dst = dstBase;
0N/A mlib_f32 *dst_end = dst + width;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A i = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A *(mlib_s32*)dst = GET_ARGBPRE(i);
0N/A dst++;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (; dst <= dst_end - 2; dst += 2) {
0N/A i0 = tmpsxloc >> shift;
0N/A i1 = (tmpsxloc + sxinc) >> shift;
0N/A tmpsxloc += 2*sxinc;
0N/A
0N/A dd = vis_faligndata(vis_ld_u8(src + 3*i1 ), dd);
0N/A dd = vis_faligndata(vis_ld_u8(src + 3*i1 + 1), dd);
0N/A dd = vis_faligndata(vis_ld_u8(src + 3*i1 + 2), dd);
0N/A dd = vis_faligndata(maskFF, dd);
0N/A dd = vis_faligndata(vis_ld_u8(src + 3*i0 ), dd);
0N/A dd = vis_faligndata(vis_ld_u8(src + 3*i0 + 1), dd);
0N/A dd = vis_faligndata(vis_ld_u8(src + 3*i0 + 2), dd);
0N/A dd = vis_faligndata(maskFF, dd);
0N/A
0N/A *(mlib_d64*)dst = dd;
0N/A }
0N/A
0N/A for (; dst < dst_end; dst++) {
0N/A i = tmpsxloc >> shift;
0N/A tmpsxloc += sxinc;
0N/A *(mlib_s32*)dst = GET_ARGBPRE(i);
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ByteIndexedToIntArgbPreConvert)(BLIT_PARAMS)
0N/A{
0N/A jint *pixLut = pSrcInfo->lutBase;
0N/A mlib_s32 buff[256];
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 i, i0, j;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A mlib_s32 a, r, g, b;
0N/A mlib_u32 x = pixLut[src[i]];
0N/A b = x & 0xff;
0N/A g = (x >> 8) & 0xff;
0N/A r = (x >> 16) & 0xff;
0N/A a = x >> 24;
0N/A r = mul8table[a][r];
0N/A g = mul8table[a][g];
0N/A b = mul8table[a][b];
0N/A dst[i] = (a << 24) | (r << 16) | (g << 8) | b;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A return;
0N/A }
0N/A
0N/A if (srcScan == width && dstScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A ADD_SUFF(IntArgbToIntArgbPreConvert)(pixLut, buff, 256, 1,
0N/A pSrcInfo, pDstInfo, pPrim, pCompInfo);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A dst[i] = buff[src[i]];
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A *(mlib_d64*)(dst + i) = LOAD_2F32(buff, src[i], src[i + 1]);
0N/A }
0N/A
0N/A for (; i < width; i++) {
0N/A dst[i] = buff[src[i]];
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ByteIndexedToIntArgbPreScaleConvert)(SCALE_PARAMS)
0N/A{
0N/A jint *pixLut = pSrcInfo->lutBase;
0N/A mlib_s32 buff[256];
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_s32 i, j;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A mlib_s32 a, r, g, b;
0N/A mlib_u32 x = pixLut[src[tmpsxloc >> shift]];
0N/A tmpsxloc += sxinc;
0N/A b = x & 0xff;
0N/A g = (x >> 8) & 0xff;
0N/A r = (x >> 16) & 0xff;
0N/A a = x >> 24;
0N/A r = mul8table[a][r];
0N/A g = mul8table[a][g];
0N/A b = mul8table[a][b];
0N/A dst[i] = (a << 24) | (r << 16) | (g << 8) | b;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A return;
0N/A }
0N/A
0N/A ADD_SUFF(IntArgbToIntArgbPreConvert)(pixLut, buff, 256, 1,
0N/A pSrcInfo, pDstInfo, pPrim, pCompInfo);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A mlib_s32 *dst_end = dst + width;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A *dst++ = buff[src[tmpsxloc >> shift]];
0N/A tmpsxloc += sxinc;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (; dst <= dst_end - 2; dst += 2) {
0N/A *(mlib_d64*)dst = LOAD_2F32(buff, src[tmpsxloc >> shift],
0N/A src[(tmpsxloc + sxinc) >> shift]);
0N/A tmpsxloc += 2*sxinc;
0N/A }
0N/A
0N/A for (; dst < dst_end; dst++) {
0N/A *dst = buff[src[tmpsxloc >> shift]];
0N/A tmpsxloc += sxinc;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ByteIndexedBmToIntArgbPreXparOver)(BLIT_PARAMS)
0N/A{
0N/A jint *pixLut = pSrcInfo->lutBase;
0N/A mlib_s32 buff[256];
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 dd, dzero;
0N/A mlib_s32 i, i0, j, x, mask;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A mlib_s32 a, r, g, b;
0N/A mlib_s32 x = pixLut[src[i]];
0N/A if (x < 0) {
0N/A b = x & 0xff;
0N/A g = (x >> 8) & 0xff;
0N/A r = (x >> 16) & 0xff;
0N/A a = (mlib_u32)x >> 24;
0N/A r = mul8table[a][r];
0N/A g = mul8table[a][g];
0N/A b = mul8table[a][b];
0N/A dst[i] = (a << 24) | (r << 16) | (g << 8) | b;
0N/A }
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A return;
0N/A }
0N/A
0N/A if (srcScan == width && dstScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A ADD_SUFF(IntArgbToIntArgbPreConvert)(pixLut, buff, 256, 1,
0N/A pSrcInfo, pDstInfo, pPrim, pCompInfo);
0N/A
0N/A dzero = vis_fzero();
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A
0N/A i = i0 = 0;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A x = buff[src[i]];
0N/A if (x < 0) {
0N/A dst[i] = x;
0N/A }
0N/A i0 = 1;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
0N/A dd = vis_freg_pair(((mlib_f32*)buff)[src[i]],
0N/A ((mlib_f32*)buff)[src[i + 1]]);
0N/A mask = vis_fcmplt32(dd, dzero);
0N/A vis_pst_32(dd, dst + i, mask);
0N/A }
0N/A
0N/A for (; i < width; i++) {
0N/A x = buff[src[i]];
0N/A if (x < 0) {
0N/A dst[i] = x;
0N/A }
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ByteIndexedBmToIntArgbPreScaleXparOver)(SCALE_PARAMS)
0N/A{
0N/A jint *pixLut = pSrcInfo->lutBase;
0N/A mlib_s32 buff[256];
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 dd, dzero;
0N/A mlib_s32 i, j, x, mask;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A mlib_s32 a, r, g, b;
0N/A mlib_s32 x = pixLut[src[tmpsxloc >> shift]];
0N/A tmpsxloc += sxinc;
0N/A if (x < 0) {
0N/A b = x & 0xff;
0N/A g = (x >> 8) & 0xff;
0N/A r = (x >> 16) & 0xff;
0N/A a = (mlib_u32)x >> 24;
0N/A r = mul8table[a][r];
0N/A g = mul8table[a][g];
0N/A b = mul8table[a][b];
0N/A dst[i] = (a << 24) | (r << 16) | (g << 8) | b;
0N/A }
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A return;
0N/A }
0N/A
0N/A ADD_SUFF(IntArgbToIntArgbPreConvert)(pixLut, buff, 256, 1,
0N/A pSrcInfo, pDstInfo, pPrim, pCompInfo);
0N/A
0N/A dzero = vis_fzero();
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A mlib_s32 *dst_end = dst + width;
0N/A mlib_s32 tmpsxloc = sxloc;
0N/A
0N/A PTR_ADD(src, (syloc >> shift) * srcScan);
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A x = buff[src[tmpsxloc >> shift]];
0N/A tmpsxloc += sxinc;
0N/A if (x < 0) {
0N/A *dst = x;
0N/A }
0N/A dst++;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (; dst <= dst_end - 2; dst += 2) {
0N/A dd = LOAD_2F32(buff, src[tmpsxloc >> shift],
0N/A src[(tmpsxloc + sxinc) >> shift]);
0N/A tmpsxloc += 2*sxinc;
0N/A mask = vis_fcmplt32(dd, dzero);
0N/A vis_pst_32(dd, dst, mask);
0N/A }
0N/A
0N/A for (; dst < dst_end; dst++) {
0N/A x = buff[src[tmpsxloc >> shift]];
0N/A tmpsxloc += sxinc;
0N/A if (x < 0) {
0N/A *dst = x;
0N/A }
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A syloc += syinc;
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(ByteIndexedBmToIntArgbPreXparBgCopy)(BCOPY_PARAMS)
0N/A{
0N/A jint *pixLut = pSrcInfo->lutBase;
0N/A mlib_s32 buff[256];
0N/A mlib_s32 dstScan = pDstInfo->scanStride;
0N/A mlib_s32 srcScan = pSrcInfo->scanStride;
0N/A mlib_d64 dd, dzero, d_bgpixel;
0N/A mlib_s32 i, j, x, mask;
0N/A
0N/A if (width < 16) {
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A x = pixLut[src[i]];
0N/A if (x < 0) {
0N/A mlib_s32 a, r, g, b;
0N/A b = x & 0xff;
0N/A g = (x >> 8) & 0xff;
0N/A r = (x >> 16) & 0xff;
0N/A a = (mlib_u32)x >> 24;
0N/A r = mul8table[a][r];
0N/A g = mul8table[a][g];
0N/A b = mul8table[a][b];
0N/A dst[i] = (a << 24) | (r << 16) | (g << 8) | b;
0N/A } else {
0N/A dst[i] = bgpixel;
0N/A }
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A return;
0N/A }
0N/A
0N/A ADD_SUFF(IntArgbToIntArgbPreConvert)(pixLut, buff, 256, 1,
0N/A pSrcInfo, pDstInfo, pPrim, pCompInfo);
0N/A
0N/A if (srcScan == width && dstScan == 4*width) {
0N/A width *= height;
0N/A height = 1;
0N/A }
0N/A
0N/A dzero = vis_fzero();
0N/A d_bgpixel = vis_to_double_dup(bgpixel);
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = srcBase;
0N/A mlib_s32 *dst = dstBase;
0N/A mlib_s32 *dst_end;
0N/A
0N/A dst_end = dst + width;
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A x = buff[*src++];
0N/A if (x < 0) {
0N/A *dst = x;
0N/A } else {
0N/A *dst = bgpixel;
0N/A }
0N/A dst++;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (; dst <= (dst_end - 2); dst += 2) {
0N/A dd = vis_freg_pair(((mlib_f32*)buff)[src[0]],
0N/A ((mlib_f32*)buff)[src[1]]);
0N/A mask = vis_fcmplt32(dd, dzero);
0N/A *(mlib_d64*)dst = d_bgpixel;
0N/A vis_pst_32(dd, dst, mask);
0N/A src += 2;
0N/A }
0N/A
0N/A while (dst < dst_end) {
0N/A x = buff[*src++];
0N/A if (x < 0) {
0N/A *dst = x;
0N/A } else {
0N/A *dst = bgpixel;
0N/A }
0N/A dst++;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, dstScan);
0N/A PTR_ADD(srcBase, srcScan);
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/Avoid ADD_SUFF(IntArgbPreDrawGlyphListAA)(SurfaceDataRasInfo * pRasInfo,
0N/A ImageRef *glyphs,
0N/A jint totalGlyphs,
0N/A jint fgpixel, jint argbcolor,
0N/A jint clipLeft, jint clipTop,
0N/A jint clipRight, jint clipBottom,
0N/A NativePrimitive * pPrim,
0N/A CompositeInfo * pCompInfo)
0N/A{
0N/A mlib_s32 glyphCounter;
0N/A mlib_s32 scan = pRasInfo->scanStride;
0N/A mlib_u8 *dstBase, *dstBase0;
0N/A mlib_s32 i, j;
0N/A mlib_d64 dmix0, dmix1, dd, d0, d1, e0, e1;
0N/A mlib_d64 done, d_half;
0N/A mlib_s32 pix;
0N/A mlib_f32 srcG_f;
0N/A
0N/A done = vis_to_double_dup(0x7fff7fff);
0N/A d_half = vis_to_double_dup((1 << (16 + 6)) | (1 << 6));
0N/A
0N/A srcG_f = vis_to_float(argbcolor);
0N/A
0N/A for (glyphCounter = 0; glyphCounter < totalGlyphs; glyphCounter++) {
0N/A const jubyte *pixels, *pixels0;
0N/A unsigned int rowBytes;
0N/A int left, top;
0N/A int width, height;
0N/A int right, bottom;
0N/A
0N/A pixels = (const jubyte *) glyphs[glyphCounter].pixels;
0N/A
0N/A if (!pixels) continue;
0N/A
0N/A left = glyphs[glyphCounter].x;
0N/A top = glyphs[glyphCounter].y;
0N/A width = glyphs[glyphCounter].width;
0N/A height = glyphs[glyphCounter].height;
0N/A rowBytes = width;
0N/A right = left + width;
0N/A bottom = top + height;
0N/A if (left < clipLeft) {
0N/A pixels += clipLeft - left;
0N/A left = clipLeft;
0N/A }
0N/A if (top < clipTop) {
0N/A pixels += (clipTop - top) * rowBytes;
0N/A top = clipTop;
0N/A }
0N/A if (right > clipRight) {
0N/A right = clipRight;
0N/A }
0N/A if (bottom > clipBottom) {
0N/A bottom = clipBottom;
0N/A }
0N/A if (right <= left || bottom <= top) {
0N/A continue;
0N/A }
0N/A width = right - left;
0N/A height = bottom - top;
0N/A
0N/A dstBase = pRasInfo->rasBase;
0N/A PTR_ADD(dstBase, top*scan + 4*left);
0N/A
0N/A pixels0 = pixels;
0N/A dstBase0 = dstBase;
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = (void*)pixels;
0N/A mlib_s32 *dst, *dst_end;
0N/A
0N/A dst = (void*)dstBase;
0N/A dst_end = dst + width;
0N/A
0N/A ADD_SUFF(IntArgbPreToIntArgbConvert)(dstBase, dstBase, width, 1,
0N/A pRasInfo, pRasInfo,
0N/A pPrim, pCompInfo);
0N/A
0N/A vis_write_gsr(0 << 3);
0N/A
0N/A if ((mlib_s32)dst & 7) {
0N/A pix = *src++;
0N/A dd = vis_fpadd16(MUL8_VIS(srcG_f, pix), d_half);
0N/A dd = vis_fpadd16(MUL8_VIS(*(mlib_f32*)dst, 255 - pix), dd);
0N/A *(mlib_f32*)dst = vis_fpack16(dd);
0N/A dst++;
0N/A }
0N/A
0N/A#pragma pipeloop(0)
0N/A for (; dst <= (dst_end - 2); dst += 2) {
0N/A dmix0 = vis_freg_pair(((mlib_f32 *)vis_mul8s_tbl)[src[0]],
0N/A ((mlib_f32 *)vis_mul8s_tbl)[src[1]]);
0N/A dmix1 = vis_fpsub16(done, dmix0);
0N/A src += 2;
0N/A
0N/A dd = *(mlib_d64*)dst;
0N/A d0 = vis_fmul8x16al(srcG_f, vis_read_hi(dmix0));
0N/A d1 = vis_fmul8x16al(srcG_f, vis_read_lo(dmix0));
0N/A e0 = vis_fmul8x16al(vis_read_hi(dd), vis_read_hi(dmix1));
0N/A e1 = vis_fmul8x16al(vis_read_lo(dd), vis_read_lo(dmix1));
0N/A d0 = vis_fpadd16(vis_fpadd16(d0, d_half), e0);
0N/A d1 = vis_fpadd16(vis_fpadd16(d1, d_half), e1);
0N/A dd = vis_fpack16_pair(d0, d1);
0N/A
0N/A *(mlib_d64*)dst = dd;
0N/A }
0N/A
0N/A while (dst < dst_end) {
0N/A pix = *src++;
0N/A dd = vis_fpadd16(MUL8_VIS(srcG_f, pix), d_half);
0N/A dd = vis_fpadd16(MUL8_VIS(*(mlib_f32*)dst, 255 - pix), dd);
0N/A *(mlib_f32*)dst = vis_fpack16(dd);
0N/A dst++;
0N/A }
0N/A
0N/A PTR_ADD(dstBase, scan);
0N/A pixels += rowBytes;
0N/A }
0N/A
0N/A pixels = pixels0;
0N/A dstBase = dstBase0;
0N/A
0N/A for (j = 0; j < height; j++) {
0N/A mlib_u8 *src = (void*)pixels;
0N/A mlib_s32 *dst = (void*)dstBase;
0N/A
0N/A for (i = 0; i < width; i++) {
0N/A if (src[i] == 255) dst[i] = fgpixel;
0N/A }
0N/A PTR_ADD(dstBase, scan);
0N/A pixels += rowBytes;
0N/A }
0N/A }
0N/A}
0N/A
0N/A/***************************************************************/
0N/A
0N/A#endif /* JAVA2D_NO_MLIB */