|
Java example source code file (vis_IntBgr.c)
The vis_IntBgr.c Java example source code/* * 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_AlphaMacros.h" /***************************************************************/ #define ARGB_to_GBGR(x) \ (x << 16) | (x & 0xff00) | ((x >> 16) & 0xff) /***************************************************************/ #define ARGB_to_BGR(x) \ ((x << 16) & 0xff0000) | (x & 0xff00) | ((x >> 16) & 0xff) /***************************************************************/ #define READ_Bgr(i) \ (src[3*i] << 16) | (src[3*i + 1] << 8) | src[3*i + 2] /***************************************************************/ #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)); \ } /***************************************************************/ #define ARGB_to_BGR_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_fzeros(), vis_read_lo(t0)); \ dst = vis_fpmerge(vis_read_hi(t2), vis_read_lo(t1)); \ } /***************************************************************/ void ADD_SUFF(IntBgrToIntArgbConvert)(BLIT_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd, amask; mlib_s32 i, i0, j, x; if (dstScan == 4*width && srcScan == 4*width) { width *= height; height = 1; } amask = vis_to_double_dup(0xFF000000); vis_alignaddr(NULL, 7); for (j = 0; j < height; j++) { mlib_u32 *src = srcBase; mlib_u32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { x = src[i]; dst[i] = 0xff000000 | ARGB_to_GBGR(x); i0 = 1; } #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 2; i += 2) { ARGB2ABGR_FL2(dd, ((mlib_f32*)src)[i], ((mlib_f32*)src)[i + 1]); *(mlib_d64*)(dst + i) = vis_for(dd, amask); } if (i < width) { x = src[i]; dst[i] = 0xff000000 | ARGB_to_GBGR(x); } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(IntBgrToIntArgbScaleConvert)(SCALE_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd, amask; mlib_s32 j, x; amask = vis_to_double_dup(0xFF000000); vis_alignaddr(NULL, 7); for (j = 0; j < height; j++) { mlib_u32 *src = srcBase; mlib_u32 *dst = dstBase; mlib_u32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); if ((mlib_s32)dst & 7) { x = src[tmpsxloc >> shift]; *dst++ = 0xff000000 | ARGB_to_GBGR(x); tmpsxloc += sxinc; } #pragma pipeloop(0) for (; dst <= dst_end - 2; dst += 2) { ARGB2ABGR_FL2(dd, ((mlib_f32*)src)[tmpsxloc >> shift], ((mlib_f32*)src)[(tmpsxloc + sxinc) >> shift]); *(mlib_d64*)dst = vis_for(dd, amask); tmpsxloc += 2*sxinc; } for (; dst < dst_end; dst++) { x = src[tmpsxloc >> shift]; *dst++ = 0xff000000 | ARGB_to_GBGR(x); tmpsxloc += sxinc; } PTR_ADD(dstBase, dstScan); syloc += syinc; } } /***************************************************************/ void ADD_SUFF(IntArgbToIntBgrConvert)(BLIT_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 i, i0, j, x; if (dstScan == 4*width && srcScan == 4*width) { width *= height; height = 1; } for (j = 0; j < height; j++) { mlib_u32 *src = srcBase; mlib_u32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { x = src[i]; dst[i] = ARGB_to_GBGR(x); i0 = 1; } #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 2; i += 2) { ARGB_to_GBGR_FL2(dd, ((mlib_f32*)src)[i], ((mlib_f32*)src)[i + 1]); *(mlib_d64*)(dst + i) = dd; } if (i < width) { x = src[i]; dst[i] = ARGB_to_GBGR(x); } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(IntArgbToIntBgrScaleConvert)(SCALE_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 j, x; for (j = 0; j < height; j++) { mlib_u32 *src = srcBase; mlib_u32 *dst = dstBase; mlib_u32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); if ((mlib_s32)dst & 7) { x = src[tmpsxloc >> shift]; *dst++ = ARGB_to_GBGR(x); tmpsxloc += sxinc; } #pragma pipeloop(0) for (; dst <= dst_end - 2; dst += 2) { ARGB_to_GBGR_FL2(dd, ((mlib_f32*)src)[tmpsxloc >> shift], ((mlib_f32*)src)[(tmpsxloc + sxinc) >> shift]); *(mlib_d64*)dst = dd; tmpsxloc += 2*sxinc; } for (; dst < dst_end; dst++) { x = src[tmpsxloc >> shift]; *dst++ = ARGB_to_GBGR(x); tmpsxloc += sxinc; } PTR_ADD(dstBase, dstScan); syloc += syinc; } } /***************************************************************/ #define INSERT_U8_34R { \ mlib_d64 sda, sdb, sdc, sdd; \ mlib_d64 sde, sdf, sdg, sdh; \ mlib_d64 sdi, sdj, sdk, sdl; \ mlib_d64 sdm; \ \ sda = vis_fpmerge(vis_read_hi(sd0), vis_read_lo(sd1)); \ sdb = vis_fpmerge(vis_read_lo(sd0), vis_read_hi(sd2)); \ sdc = vis_fpmerge(vis_read_hi(sd1), vis_read_lo(sd2)); \ sdd = vis_fpmerge(vis_read_hi(sda), vis_read_lo(sdb)); \ sde = vis_fpmerge(vis_read_lo(sda), vis_read_hi(sdc)); \ sdf = vis_fpmerge(vis_read_hi(sdb), vis_read_lo(sdc)); \ sdg = vis_fpmerge(vis_read_hi(sdd), vis_read_lo(sde)); \ sdh = vis_fpmerge(vis_read_lo(sdd), vis_read_hi(sdf)); \ sdi = vis_fpmerge(vis_read_hi(sde), vis_read_lo(sdf)); \ sdj = vis_fpmerge(vis_read_hi(sdg), vis_read_hi(sdi)); \ sdk = vis_fpmerge(vis_read_lo(sdg), vis_read_lo(sdi)); \ sdl = vis_fpmerge(vis_read_hi(sFF), vis_read_hi(sdh)); \ sdm = vis_fpmerge(vis_read_lo(sFF), vis_read_lo(sdh)); \ dd0 = vis_fpmerge(vis_read_hi(sdl), vis_read_hi(sdj)); \ dd1 = vis_fpmerge(vis_read_lo(sdl), vis_read_lo(sdj)); \ dd2 = vis_fpmerge(vis_read_hi(sdm), vis_read_hi(sdk)); \ dd3 = vis_fpmerge(vis_read_lo(sdm), vis_read_lo(sdk)); \ } /***************************************************************/ void ADD_SUFF(ThreeByteBgrToIntBgrConvert)(BLIT_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 *sp; mlib_d64 sFF; mlib_d64 s0, s1, s2, s3, sd0, sd1, sd2, dd0, dd1, dd2, dd3; mlib_s32 i, i0, j; if (width < 16) { for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_u32 *dst = dstBase; for (i = 0; i < width; i++) { dst[i] = READ_Bgr(i); } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } return; } if (srcScan == 3*width && dstScan == 4*width) { width *= height; height = 1; } sFF = vis_fzero(); for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_f32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { ((mlib_s32*)dst)[i] = READ_Bgr(i); i0 = 1; } sp = vis_alignaddr(src, 3*i0); s3 = *sp++; #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 8; i += 8) { s0 = s3; s1 = *sp++; s2 = *sp++; s3 = *sp++; sd0 = vis_faligndata(s0, s1); sd1 = vis_faligndata(s1, s2); sd2 = vis_faligndata(s2, s3); INSERT_U8_34R *(mlib_d64*)(dst + i ) = dd0; *(mlib_d64*)(dst + i + 2) = dd1; *(mlib_d64*)(dst + i + 4) = dd2; *(mlib_d64*)(dst + i + 6) = dd3; } for (; i < width; i++) { ((mlib_s32*)dst)[i] = READ_Bgr(i); } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(ThreeByteBgrToIntBgrScaleConvert)(SCALE_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd, dzero; mlib_s32 i, i0, i1, j; if (width < 16) { for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_s32 *dst = dstBase; mlib_s32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); for (; dst < dst_end; dst++) { i = tmpsxloc >> shift; tmpsxloc += sxinc; *(mlib_s32*)dst = READ_Bgr(i); } PTR_ADD(dstBase, dstScan); syloc += syinc; } return; } dzero = vis_fzero(); vis_alignaddr(NULL, 7); for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_f32 *dst = dstBase; mlib_f32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); if ((mlib_s32)dst & 7) { i = tmpsxloc >> shift; tmpsxloc += sxinc; *(mlib_s32*)dst = READ_Bgr(i); dst++; } #pragma pipeloop(0) for (; dst <= dst_end - 2; dst += 2) { i0 = tmpsxloc >> shift; i1 = (tmpsxloc + sxinc) >> shift; tmpsxloc += 2*sxinc; dd = vis_faligndata(vis_ld_u8(src + 3*i1 + 2), dd); dd = vis_faligndata(vis_ld_u8(src + 3*i1 + 1), dd); dd = vis_faligndata(vis_ld_u8(src + 3*i1 ), dd); dd = vis_faligndata(dzero, dd); dd = vis_faligndata(vis_ld_u8(src + 3*i0 + 2), dd); dd = vis_faligndata(vis_ld_u8(src + 3*i0 + 1), dd); dd = vis_faligndata(vis_ld_u8(src + 3*i0 ), dd); dd = vis_faligndata(dzero, dd); *(mlib_d64*)dst = dd; } for (; dst < dst_end; dst++) { i = tmpsxloc >> shift; tmpsxloc += sxinc; *(mlib_s32*)dst = READ_Bgr(i); } PTR_ADD(dstBase, dstScan); syloc += syinc; } } /***************************************************************/ void ADD_SUFF(IntArgbBmToIntBgrXparOver)(BLIT_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 i, i0, j, mask, x; if (dstScan == 4*width && srcScan == 4*width) { width *= height; height = 1; } for (j = 0; j < height; j++) { mlib_s32 *src = srcBase; mlib_s32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { if (*(mlib_u8*)(src + i)) { x = src[i]; dst[i] = ARGB_to_GBGR(x); } i0 = 1; } #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 2; i += 2) { ARGB_to_GBGR_FL2(dd, ((mlib_f32*)src)[i], ((mlib_f32*)src)[i + 1]); mask = (((-*(mlib_u8*)(src + i)) >> 31) & 2) | (((-*(mlib_u8*)(src + i + 1)) >> 31) & 1); vis_pst_32(dd, dst + i, mask); } if (i < width) { if (*(mlib_u8*)(src + i)) { x = src[i]; dst[i] = ARGB_to_GBGR(x); } } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(IntArgbBmToIntBgrScaleXparOver)(SCALE_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 j, mask; for (j = 0; j < height; j++) { mlib_s32 *src = srcBase; mlib_s32 *dst = dstBase; mlib_s32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); if ((mlib_s32)dst & 7) { mlib_s32 *pp = src + (tmpsxloc >> shift); if (*(mlib_u8*)pp) { *dst = ARGB_to_GBGR(*pp); } dst++; tmpsxloc += sxinc; } #pragma pipeloop(0) for (; dst <= dst_end - 2; dst += 2) { mlib_s32 *pp0 = src + (tmpsxloc >> shift); mlib_s32 *pp1 = src + ((tmpsxloc + sxinc) >> shift); ARGB_to_GBGR_FL2(dd, *(mlib_f32*)pp0, *(mlib_f32*)pp1); mask = (((-*(mlib_u8*)pp0) >> 31) & 2) | ((mlib_u32)(-*(mlib_u8*)pp1) >> 31); vis_pst_32(dd, dst, mask); tmpsxloc += 2*sxinc; } for (; dst < dst_end; dst++) { mlib_s32 *pp = src + (tmpsxloc >> shift); if (*(mlib_u8*)pp) { *dst = ARGB_to_GBGR(*pp); } } PTR_ADD(dstBase, dstScan); syloc += syinc; } } /***************************************************************/ void ADD_SUFF(IntArgbBmToIntBgrXparBgCopy)(BCOPY_PARAMS) { mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd, d_bgpixel; mlib_s32 i, i0, j, mask; if (dstScan == 4*width && srcScan == 4*width) { width *= height; height = 1; } vis_alignaddr(NULL, 1); d_bgpixel = vis_to_double_dup(bgpixel); for (j = 0; j < height; j++) { mlib_s32 *src = srcBase; mlib_s32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { if (*(mlib_u8*)(src + i)) { dst[i] = ARGB_to_GBGR(src[i]); } else { dst[i] = bgpixel; } i0 = 1; } #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 2; i += 2) { ARGB_to_GBGR_FL2(dd, ((mlib_f32*)src)[i], ((mlib_f32*)src)[i + 1]); mask = (((-*(mlib_u8*)(src + i)) >> 31) & 2) | (((-*(mlib_u8*)(src + i + 1)) >> 31) & 1); *(mlib_d64*)(dst + i) = d_bgpixel; vis_pst_32(dd, dst + i, mask); } if (i < width) { if (*(mlib_u8*)(src + i)) { dst[i] = ARGB_to_GBGR(src[i]); } else { dst[i] = bgpixel; } } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(ByteIndexedToIntBgrConvert)(BLIT_PARAMS) { jint *pixLut = pSrcInfo->lutBase; mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 i, i0, j, x; if (srcScan == width && dstScan == 4*width) { width *= height; height = 1; } for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_s32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { x = pixLut[src[i]]; dst[i] = ARGB_to_GBGR(x); i0 = 1; } #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 2; i += 2) { ARGB_to_GBGR_FL2(dd, ((mlib_f32*)pixLut)[src[i]], ((mlib_f32*)pixLut)[src[i + 1]]); *(mlib_d64*)(dst + i) = dd; } for (; i < width; i++) { x = pixLut[src[i]]; dst[i] = ARGB_to_GBGR(x); } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(ByteIndexedToIntBgrScaleConvert)(SCALE_PARAMS) { jint *pixLut = pSrcInfo->lutBase; mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 j, x; for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_s32 *dst = dstBase; mlib_s32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); if ((mlib_s32)dst & 7) { x = pixLut[src[tmpsxloc >> shift]]; *dst++ = ARGB_to_GBGR(x); tmpsxloc += sxinc; } #pragma pipeloop(0) for (; dst <= dst_end - 2; dst += 2) { mlib_f32 f0 = ((mlib_f32*)pixLut)[src[tmpsxloc >> shift]]; mlib_f32 f1 = ((mlib_f32*)pixLut)[src[(tmpsxloc + sxinc) >> shift]]; ARGB_to_GBGR_FL2(dd, f0, f1); *(mlib_d64*)dst = dd; tmpsxloc += 2*sxinc; } for (; dst < dst_end; dst++) { x = pixLut[src[tmpsxloc >> shift]]; *dst++ = ARGB_to_GBGR(x); tmpsxloc += sxinc; } PTR_ADD(dstBase, dstScan); syloc += syinc; } } /***************************************************************/ void ADD_SUFF(ByteIndexedBmToIntBgrXparOver)(BLIT_PARAMS) { jint *pixLut = pSrcInfo->lutBase; mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 i, i0, j, x, mask; if (srcScan == width && dstScan == 4*width) { width *= height; height = 1; } for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_s32 *dst = dstBase; i = i0 = 0; if ((mlib_s32)dst & 7) { x = pixLut[src[i]]; if (x < 0) { dst[i] = ARGB_to_BGR(x); } i0 = 1; } #pragma pipeloop(0) for (i = i0; i <= (mlib_s32)width - 2; i += 2) { mlib_f32 *pp0 = (mlib_f32*)pixLut + src[i]; mlib_f32 *pp1 = (mlib_f32*)pixLut + src[i + 1]; ARGB_to_BGR_FL2(dd, *pp0, *pp1); mask = (((*(mlib_u8*)pp0) >> 6) & 2) | ((*(mlib_u8*)pp1) >> 7); vis_pst_32(dd, dst + i, mask); } for (; i < width; i++) { x = pixLut[src[i]]; if (x < 0) { dst[i] = ARGB_to_BGR(x); } } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(ByteIndexedBmToIntBgrScaleXparOver)(SCALE_PARAMS) { jint *pixLut = pSrcInfo->lutBase; mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd; mlib_s32 j, x, mask; for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_s32 *dst = dstBase; mlib_s32 *dst_end = dst + width; mlib_s32 tmpsxloc = sxloc; PTR_ADD(src, (syloc >> shift) * srcScan); if ((mlib_s32)dst & 7) { x = pixLut[src[tmpsxloc >> shift]]; tmpsxloc += sxinc; if (x < 0) { *dst = ARGB_to_BGR(x); } dst++; } #pragma pipeloop(0) for (; dst <= dst_end - 2; dst += 2) { mlib_f32 *p0 = (mlib_f32*)pixLut + src[tmpsxloc >> shift]; mlib_f32 *p1 = (mlib_f32*)pixLut + src[(tmpsxloc + sxinc) >> shift]; ARGB_to_BGR_FL2(dd, *p0, *p1); mask = (((*(mlib_u8*)p0) >> 6) & 2) | ((*(mlib_u8*)p1) >> 7); tmpsxloc += 2*sxinc; vis_pst_32(dd, dst, mask); } for (; dst < dst_end; dst++) { x = pixLut[src[tmpsxloc >> shift]]; tmpsxloc += sxinc; if (x < 0) { *dst = ARGB_to_BGR(x); } } PTR_ADD(dstBase, dstScan); syloc += syinc; } } /***************************************************************/ void ADD_SUFF(ByteIndexedBmToIntBgrXparBgCopy)(BCOPY_PARAMS) { jint *pixLut = pSrcInfo->lutBase; mlib_s32 dstScan = pDstInfo->scanStride; mlib_s32 srcScan = pSrcInfo->scanStride; mlib_d64 dd, d_bgpixel; mlib_s32 j, x, mask; if (srcScan == width && dstScan == 4*width) { width *= height; height = 1; } d_bgpixel = vis_to_double_dup(bgpixel); for (j = 0; j < height; j++) { mlib_u8 *src = srcBase; mlib_s32 *dst = dstBase; mlib_s32 *dst_end; dst_end = dst + width; if ((mlib_s32)dst & 7) { x = pixLut[*src++]; if (x < 0) { *dst = ARGB_to_GBGR(x); } else { *dst = bgpixel; } dst++; } #pragma pipeloop(0) for (; dst <= (dst_end - 2); dst += 2) { mlib_f32 *pp0 = (mlib_f32*)pixLut + src[0]; mlib_f32 *pp1 = (mlib_f32*)pixLut + src[1]; ARGB_to_GBGR_FL2(dd, *pp0, *pp1); mask = (((*(mlib_u8*)pp0) >> 6) & 2) | ((*(mlib_u8*)pp1) >> 7); *(mlib_d64*)dst = d_bgpixel; vis_pst_32(dd, dst, mask); src += 2; } while (dst < dst_end) { x = pixLut[*src++]; if (x < 0) { *dst = ARGB_to_GBGR(x); } else { *dst = bgpixel; } dst++; } PTR_ADD(dstBase, dstScan); PTR_ADD(srcBase, srcScan); } } /***************************************************************/ void ADD_SUFF(IntBgrDrawGlyphListAA)(GLYPH_LIST_PARAMS) { mlib_s32 glyphCounter; mlib_s32 scan = pRasInfo->scanStride; mlib_u8 *dstBase; mlib_s32 j; mlib_d64 dmix0, dmix1, dd, d0, d1, e0, e1, fgpixel_d; mlib_d64 done, done16, d_half, maskRGB, dzero; mlib_s32 pix, mask, mask_z; mlib_f32 srcG_f; done = vis_to_double_dup(0x7fff7fff); done16 = vis_to_double_dup(0x7fff); d_half = vis_to_double_dup((1 << (16 + 6)) | (1 << 6)); fgpixel_d = vis_to_double_dup(fgpixel); srcG_f = vis_to_float(argbcolor); maskRGB = vis_to_double_dup(0xffffff); dzero = vis_fzero(); ARGB2ABGR_FL(srcG_f) vis_write_gsr(0 << 3); for (glyphCounter = 0; glyphCounter < totalGlyphs; glyphCounter++) { const jubyte *pixels; unsigned int rowBytes; int left, top; int width, height; int right, bottom; pixels = (const jubyte *) glyphs[glyphCounter].pixels; if (!pixels) continue; left = glyphs[glyphCounter].x; top = glyphs[glyphCounter].y; width = glyphs[glyphCounter].width; height = glyphs[glyphCounter].height; rowBytes = width; right = left + width; bottom = top + height; if (left < clipLeft) { pixels += clipLeft - left; left = clipLeft; } if (top < clipTop) { pixels += (clipTop - top) * rowBytes; top = clipTop; } if (right > clipRight) { right = clipRight; } if (bottom > clipBottom) { bottom = clipBottom; } if (right <= left || bottom <= top) { continue; } width = right - left; height = bottom - top; dstBase = pRasInfo->rasBase; PTR_ADD(dstBase, top*scan + 4*left); for (j = 0; j < height; j++) { mlib_u8 *src = (void*)pixels; mlib_s32 *dst, *dst_end; dst = (void*)dstBase; dst_end = dst + width; if ((mlib_s32)dst & 7) { pix = *src++; if (pix) { dd = vis_fpadd16(MUL8_VIS(srcG_f, pix), d_half); dd = vis_fpadd16(MUL8_VIS(*(mlib_f32*)dst, 255 - pix), dd); *(mlib_f32*)dst = vis_fands(vis_fpack16(dd), vis_read_hi(maskRGB)); if (pix == 255) *(mlib_f32*)dst = vis_read_hi(fgpixel_d); } dst++; } #pragma pipeloop(0) for (; dst <= (dst_end - 2); dst += 2) { dmix0 = vis_freg_pair(((mlib_f32 *)vis_mul8s_tbl)[src[0]], ((mlib_f32 *)vis_mul8s_tbl)[src[1]]); mask = vis_fcmplt32(dmix0, done16); mask_z = vis_fcmpne32(dmix0, dzero); dmix1 = vis_fpsub16(done, dmix0); src += 2; dd = *(mlib_d64*)dst; d0 = vis_fmul8x16al(srcG_f, vis_read_hi(dmix0)); d1 = vis_fmul8x16al(srcG_f, vis_read_lo(dmix0)); e0 = vis_fmul8x16al(vis_read_hi(dd), vis_read_hi(dmix1)); e1 = vis_fmul8x16al(vis_read_lo(dd), vis_read_lo(dmix1)); d0 = vis_fpadd16(vis_fpadd16(d0, d_half), e0); d1 = vis_fpadd16(vis_fpadd16(d1, d_half), e1); dd = vis_fpack16_pair(d0, d1); dd = vis_fand(dd, maskRGB); vis_pst_32(fgpixel_d, dst, mask_z); vis_pst_32(dd, dst, mask & mask_z); } while (dst < dst_end) { pix = *src++; if (pix) { dd = vis_fpadd16(MUL8_VIS(srcG_f, pix), d_half); dd = vis_fpadd16(MUL8_VIS(*(mlib_f32*)dst, 255 - pix), dd); *(mlib_f32*)dst = vis_fands(vis_fpack16(dd), vis_read_hi(maskRGB)); if (pix == 255) *(mlib_f32*)dst = vis_read_hi(fgpixel_d); } dst++; } PTR_ADD(dstBase, scan); pixels += rowBytes; } } } /***************************************************************/ #endif /* JAVA2D_NO_MLIB */ Other Java examples (source code examples)Here is a short list of links related to this Java vis_IntBgr.c source code file: |
... this post is sponsored by my books ... | |
#1 New Release! |
FP Best Seller |
Copyright 1998-2024 Alvin Alexander, alvinalexander.com
All Rights Reserved.
A percentage of advertising revenue from
pages under the /java/jwarehouse
URI on this website is
paid back to open source projects.