/*
* Copyright (c) 2003, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation. Oracle designates this
* particular file as subject to the "Classpath" exception as provided
* by Oracle in the LICENSE file that accompanied this code.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
#if !defined(JAVA2D_NO_MLIB) || defined(MLIB_ADD_SUFF)
#include <vis_proto.h>
#include "java2d_Mlib.h"
/***************************************************************/
#define ARGB_XOR(index, chan) \
{ \
jint srcpixel = src_ptr[index]; \
jint neg_mask = srcpixel >> 31; \
dst_ptr[index] ^= (srcpixel ^ xorpixel) & (neg_mask &~ alphamask); \
}
/***************************************************************/
#define BGR_XOR(index, chan) \
{ \
jint srcpixel = src_ptr[index]; \
jint neg_mask = srcpixel >> 31; \
srcpixel = (srcpixel << 16) | (srcpixel & 0xff00) | \
((srcpixel >> 16) & 0xff); \
dst_ptr[index] ^= (srcpixel ^ xorpixel) & (neg_mask &~ alphamask); \
}
/***************************************************************/
#define ARGB_BM_XOR(index, chan) \
{ \
jint srcpixel = src_ptr[index]; \
jint neg_mask = srcpixel >> 31; \
srcpixel |= 0xFF000000; \
dst_ptr[index] ^= (srcpixel ^ xorpixel) & (neg_mask &~ alphamask); \
}
/***************************************************************/
#define RGBX_XOR(index, chan) \
{ \
jint srcpixel = src_ptr[index]; \
jint neg_mask = srcpixel >> 31; \
dst_ptr[index] ^= ((srcpixel << 8) ^ xorpixel) & \
(neg_mask &~ alphamask); \
}
/***************************************************************/
#define ARGB_to_GBGR_FL2(dst, src0, src1) { \
mlib_d64 t0, t1, t2; \
t0 = vis_fpmerge(src0, src1); \
t1 = vis_fpmerge(vis_read_lo(t0), vis_read_hi(t0)); \
t2 = vis_fpmerge(vis_read_lo(t0), vis_read_lo(t0)); \
dst = vis_fpmerge(vis_read_hi(t2), vis_read_lo(t1)); \
}
/***************************************************************/
#ifdef MLIB_ADD_SUFF
#pragma weak IntArgbToIntRgbXorBlit_F = IntArgbToIntArgbXorBlit_F
#else
#pragma weak IntArgbToIntRgbXorBlit = IntArgbToIntArgbXorBlit
#endif
/***************************************************************/
void ADD_SUFF(IntArgbToIntArgbXorBlit)(BLIT_PARAMS)
{
mlib_s32 dstScan = pDstInfo->scanStride;
mlib_s32 srcScan = pSrcInfo->scanStride;
mlib_s32 xorpixel = pCompInfo->details.xorPixel;
mlib_s32 alphamask = pCompInfo->alphaMask;
mlib_s32 i, j;
mlib_d64 res, xorpixel64, alphamask64, dzero;
if (width < 8) {
LOOP_DST_SRC(AnyInt, 1, dstBase, dstScan, srcBase, srcScan, ARGB_XOR);
return;
}
if (dstScan == 4*width && srcScan == dstScan) {
width *= height;
height = 1;
}
xorpixel64 = vis_to_double_dup(xorpixel);
alphamask64 = vis_to_double_dup(alphamask);
dzero = vis_fzero();
for (j = 0; j < height; j++) {
mlib_s32 *dst_ptr = dstBase;
mlib_s32 *src_ptr = srcBase;
mlib_s32 size = width;
if ((mlib_s32)dst_ptr & 7) {
ARGB_XOR(0, 0);
dst_ptr++;
src_ptr++;
size--;
}
#pragma pipeloop(0)
for (i = 0; i <= size - 2; i += 2) {
mlib_s32 neg_mask;
mlib_f32 *pp0 = (mlib_f32*)src_ptr + i;
mlib_f32 *pp1 = (mlib_f32*)src_ptr + i + 1;
neg_mask = (((*(mlib_u8*)pp0) >> 6) & 2) | ((*(mlib_u8*)pp1) >> 7);
res = vis_freg_pair(*pp0, *pp1);
res = vis_fxor(res, xorpixel64);
res = vis_fandnot(alphamask64, res);
res = vis_fxor(res, *(mlib_d64*)(dst_ptr + i));
vis_pst_32(res, dst_ptr + i, neg_mask);
}
if (i < size) {
ARGB_XOR(i, 0);
}
PTR_ADD(dstBase, dstScan);
PTR_ADD(srcBase, srcScan);
}
}
/***************************************************************/
void ADD_SUFF(IntArgbToIntBgrXorBlit)(BLIT_PARAMS)
{
mlib_s32 dstScan = pDstInfo->scanStride;
mlib_s32 srcScan = pSrcInfo->scanStride;
mlib_s32 xorpixel = pCompInfo->details.xorPixel;
mlib_s32 alphamask = pCompInfo->alphaMask;
mlib_s32 i, j;
mlib_d64 res, xorpixel64, alphamask64, dzero;
if (width < 8) {
LOOP_DST_SRC(AnyInt, 1, dstBase, dstScan, srcBase, srcScan, BGR_XOR);
return;
}
if (dstScan == 4*width && srcScan == dstScan) {
width *= height;
height = 1;
}
xorpixel64 = vis_to_double_dup(xorpixel);
alphamask64 = vis_to_double_dup(alphamask);
dzero = vis_fzero();
for (j = 0; j < height; j++) {
mlib_s32 *dst_ptr = dstBase;
mlib_s32 *src_ptr = srcBase;
mlib_s32 size = width;
if ((mlib_s32)dst_ptr & 7) {
BGR_XOR(0, 0);
dst_ptr++;
src_ptr++;
size--;
}
#pragma pipeloop(0)
for (i = 0; i <= size - 2; i += 2) {
mlib_s32 neg_mask;
mlib_f32 *pp0 = (mlib_f32*)src_ptr + i;
mlib_f32 *pp1 = (mlib_f32*)src_ptr + i + 1;
neg_mask = (((*(mlib_u8*)pp0) >> 6) & 2) | ((*(mlib_u8*)pp1) >> 7);
ARGB_to_GBGR_FL2(res, *pp0, *pp1);
res = vis_fxor(res, xorpixel64);
res = vis_fandnot(alphamask64, res);
res = vis_fxor(res, *(mlib_d64*)(dst_ptr + i));
vis_pst_32(res, dst_ptr + i, neg_mask);
}
if (i < size) {
BGR_XOR(i, 0);
}
PTR_ADD(dstBase, dstScan);
PTR_ADD(srcBase, srcScan);
}
}
/***************************************************************/
void ADD_SUFF(IntArgbToIntArgbBmXorBlit)(BLIT_PARAMS)
{
mlib_s32 dstScan = pDstInfo->scanStride;
mlib_s32 srcScan = pSrcInfo->scanStride;
mlib_s32 xorpixel = pCompInfo->details.xorPixel;
mlib_s32 alphamask = pCompInfo->alphaMask;
mlib_s32 i, j, neg_mask;
mlib_d64 res, xorpixel64, alphamask64, dzero, dFF;
if (width < 8) {
LOOP_DST_SRC(AnyInt, 1, dstBase, dstScan, srcBase, srcScan,
ARGB_BM_XOR);
return;
}
if (dstScan == 4*width && srcScan == dstScan) {
width *= height;
height = 1;
}
xorpixel64 = vis_to_double_dup(xorpixel);
alphamask64 = vis_to_double_dup(alphamask);
dzero = vis_fzero();
dFF = vis_to_double_dup(0xFF000000);
for (j = 0; j < height; j++) {
mlib_s32 *dst_ptr = dstBase;
mlib_s32 *src_ptr = srcBase;
mlib_s32 size = width;
if ((mlib_s32)dst_ptr & 7) {
ARGB_BM_XOR(0, 0);
dst_ptr++;
src_ptr++;
size--;
}
#pragma pipeloop(0)
for (i = 0; i <= size - 2; i += 2) {
mlib_s32 neg_mask;
mlib_f32 *pp0 = (mlib_f32*)src_ptr + i;
mlib_f32 *pp1 = (mlib_f32*)src_ptr + i + 1;
neg_mask = (((*(mlib_u8*)pp0) >> 6) & 2) | ((*(mlib_u8*)pp1) >> 7);
res = vis_freg_pair(*pp0, *pp1);
res = vis_for(res, dFF);
res = vis_fxor(res, xorpixel64);
res = vis_fandnot(alphamask64, res);
res = vis_fxor(res, *(mlib_d64*)(dst_ptr + i));
vis_pst_32(res, dst_ptr + i, neg_mask);
}
if (i < size) {
ARGB_BM_XOR(i, 0);
}
PTR_ADD(dstBase, dstScan);
PTR_ADD(srcBase, srcScan);
}
}
/***************************************************************/
void ADD_SUFF(IntArgbToIntRgbxXorBlit)(BLIT_PARAMS)
{
mlib_s32 dstScan = pDstInfo->scanStride;
mlib_s32 srcScan = pSrcInfo->scanStride;
mlib_s32 xorpixel = pCompInfo->details.xorPixel;
mlib_s32 alphamask = pCompInfo->alphaMask;
mlib_s32 i, j, neg_mask;
mlib_d64 res, xorpixel64, alphamask64, rgbx_mask, dzero;
if (width < 8) {
LOOP_DST_SRC(AnyInt, 1, dstBase, dstScan, srcBase, srcScan, RGBX_XOR);
return;
}
if (dstScan == 4*width && srcScan == dstScan) {
width *= height;
height = 1;
}
xorpixel64 = vis_to_double_dup(xorpixel);
alphamask64 = vis_to_double_dup(alphamask);
rgbx_mask = vis_to_double_dup(0xFFFFFF00);
dzero = vis_fzero();
vis_alignaddr(NULL, 1);
for (j = 0; j < height; j++) {
mlib_s32 *dst_ptr = dstBase;
mlib_s32 *src_ptr = srcBase;
mlib_s32 size = width;
if ((mlib_s32)dst_ptr & 7) {
RGBX_XOR(0, 0);
dst_ptr++;
src_ptr++;
size--;
}
#pragma pipeloop(0)
for (i = 0; i <= size - 2; i += 2) {
mlib_s32 neg_mask;
mlib_f32 *pp0 = (mlib_f32*)src_ptr + i;
mlib_f32 *pp1 = (mlib_f32*)src_ptr + i + 1;
neg_mask = (((*(mlib_u8*)pp0) >> 6) & 2) | ((*(mlib_u8*)pp1) >> 7);
res = vis_freg_pair(*pp0, *pp1);
res = vis_fand(vis_faligndata(res, res), rgbx_mask);
res = vis_fxor(res, xorpixel64);
res = vis_fandnot(alphamask64, res);
res = vis_fxor(res, *(mlib_d64*)(dst_ptr + i));
vis_pst_32(res, dst_ptr + i, neg_mask);
}
if (i < size) {
RGBX_XOR(i, 0);
}
PTR_ADD(dstBase, dstScan);
PTR_ADD(srcBase, srcScan);
}
}
/***************************************************************/
void ADD_SUFF(IntArgbToFourByteAbgrPreXorBlit)(BLIT_PARAMS)
{
jint xorpixel = pCompInfo->details.xorPixel;
juint alphamask = pCompInfo->alphaMask;
jint xor0, xor1, xor2, xor3;
jint mask0, mask1, mask2, mask3;
jint *pSrc = srcBase;
jubyte *pDst = dstBase;
jint srcScan = pSrcInfo->scanStride;
jint dstScan = pDstInfo->scanStride;
/**代码未完, 请加载全部代码(NowJava.com).**/