yading@10: /*
yading@10:  * rectangle filling function
yading@10:  * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
yading@10:  *
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: /**
yading@10:  * @file
yading@10:  * useful rectangle filling function
yading@10:  * @author Michael Niedermayer <michaelni@gmx.at>
yading@10:  */
yading@10: 
yading@10: #ifndef AVCODEC_RECTANGLE_H
yading@10: #define AVCODEC_RECTANGLE_H
yading@10: 
yading@10: #include "config.h"
yading@10: #include "libavutil/common.h"
yading@10: #include "libavutil/avassert.h"
yading@10: 
yading@10: /**
yading@10:  * fill a rectangle.
yading@10:  * @param h height of the rectangle, should be a constant
yading@10:  * @param w width of the rectangle, should be a constant
yading@10:  * @param size the size of val (1, 2 or 4), should be a constant
yading@10:  */
yading@10: static av_always_inline void fill_rectangle(void *vp, int w, int h, int stride, uint32_t val, int size){
yading@10:     uint8_t *p= (uint8_t*)vp;
yading@10:     av_assert2(size==1 || size==2 || size==4);
yading@10:     av_assert2(w<=4);
yading@10: 
yading@10:     w      *= size;
yading@10:     stride *= size;
yading@10: 
yading@10:     av_assert2((((long)vp)&(FFMIN(w, 8<<(HAVE_NEON|ARCH_PPC|HAVE_MMX))-1)) == 0);
yading@10:     av_assert2((stride&(w-1))==0);
yading@10:     if(w==2){
yading@10:         const uint16_t v= size==4 ? val : val*0x0101;
yading@10:         *(uint16_t*)(p + 0*stride)= v;
yading@10:         if(h==1) return;
yading@10:         *(uint16_t*)(p + 1*stride)= v;
yading@10:         if(h==2) return;
yading@10:         *(uint16_t*)(p + 2*stride)= v;
yading@10:         *(uint16_t*)(p + 3*stride)= v;
yading@10:     }else if(w==4){
yading@10:         const uint32_t v= size==4 ? val : size==2 ? val*0x00010001 : val*0x01010101;
yading@10:         *(uint32_t*)(p + 0*stride)= v;
yading@10:         if(h==1) return;
yading@10:         *(uint32_t*)(p + 1*stride)= v;
yading@10:         if(h==2) return;
yading@10:         *(uint32_t*)(p + 2*stride)= v;
yading@10:         *(uint32_t*)(p + 3*stride)= v;
yading@10:     }else if(w==8){
yading@10:     //gcc can't optimize 64bit math on x86_32
yading@10: #if HAVE_FAST_64BIT
yading@10:         const uint64_t v=  size==2 ? val*0x0001000100010001ULL : val*0x0100000001ULL;
yading@10:         *(uint64_t*)(p + 0*stride)= v;
yading@10:         if(h==1) return;
yading@10:         *(uint64_t*)(p + 1*stride)= v;
yading@10:         if(h==2) return;
yading@10:         *(uint64_t*)(p + 2*stride)= v;
yading@10:         *(uint64_t*)(p + 3*stride)= v;
yading@10:     }else if(w==16){
yading@10:         const uint64_t v= val*0x0100000001ULL;
yading@10:         *(uint64_t*)(p + 0+0*stride)= v;
yading@10:         *(uint64_t*)(p + 8+0*stride)= v;
yading@10:         *(uint64_t*)(p + 0+1*stride)= v;
yading@10:         *(uint64_t*)(p + 8+1*stride)= v;
yading@10:         if(h==2) return;
yading@10:         *(uint64_t*)(p + 0+2*stride)= v;
yading@10:         *(uint64_t*)(p + 8+2*stride)= v;
yading@10:         *(uint64_t*)(p + 0+3*stride)= v;
yading@10:         *(uint64_t*)(p + 8+3*stride)= v;
yading@10: #else
yading@10:         const uint32_t v= size==2 ? val*0x00010001 : val;
yading@10:         *(uint32_t*)(p + 0+0*stride)= v;
yading@10:         *(uint32_t*)(p + 4+0*stride)= v;
yading@10:         if(h==1) return;
yading@10:         *(uint32_t*)(p + 0+1*stride)= v;
yading@10:         *(uint32_t*)(p + 4+1*stride)= v;
yading@10:         if(h==2) return;
yading@10:         *(uint32_t*)(p + 0+2*stride)= v;
yading@10:         *(uint32_t*)(p + 4+2*stride)= v;
yading@10:         *(uint32_t*)(p + 0+3*stride)= v;
yading@10:         *(uint32_t*)(p + 4+3*stride)= v;
yading@10:     }else if(w==16){
yading@10:         *(uint32_t*)(p + 0+0*stride)= val;
yading@10:         *(uint32_t*)(p + 4+0*stride)= val;
yading@10:         *(uint32_t*)(p + 8+0*stride)= val;
yading@10:         *(uint32_t*)(p +12+0*stride)= val;
yading@10:         *(uint32_t*)(p + 0+1*stride)= val;
yading@10:         *(uint32_t*)(p + 4+1*stride)= val;
yading@10:         *(uint32_t*)(p + 8+1*stride)= val;
yading@10:         *(uint32_t*)(p +12+1*stride)= val;
yading@10:         if(h==2) return;
yading@10:         *(uint32_t*)(p + 0+2*stride)= val;
yading@10:         *(uint32_t*)(p + 4+2*stride)= val;
yading@10:         *(uint32_t*)(p + 8+2*stride)= val;
yading@10:         *(uint32_t*)(p +12+2*stride)= val;
yading@10:         *(uint32_t*)(p + 0+3*stride)= val;
yading@10:         *(uint32_t*)(p + 4+3*stride)= val;
yading@10:         *(uint32_t*)(p + 8+3*stride)= val;
yading@10:         *(uint32_t*)(p +12+3*stride)= val;
yading@10: #endif
yading@10:     }else
yading@10:         av_assert2(0);
yading@10:     av_assert2(h==4);
yading@10: }
yading@10: 
yading@10: #endif /* AVCODEC_RECTANGLE_H */