Mercurial > mplayer.hg
view libmpcodecs/vf_mirror.c @ 26625:5b89b42f6d50
Only compile and use libmpeg2 AltiVec code when AltiVec is available. The
AltiVec code needs -maltivec to compile, but then AltiVec instructions
appear in other places of the code causing MPlayer to sigill.
Somehow upstream libmpeg2 manages not to sigill under what appear to be
the same circumstances. Enlightenment welcome.
author | diego |
---|---|
date | Sat, 03 May 2008 15:23:22 +0000 |
parents | 00fff9a3b735 |
children | 0f1b5b68af32 |
line wrap: on
line source
#include <stdio.h> #include <stdlib.h> #include <string.h> #include <inttypes.h> #include "config.h" #include "mp_msg.h" #include "img_format.h" #include "mp_image.h" #include "vf.h" static void mirror(unsigned char* dst,unsigned char* src,int dststride,int srcstride,int w,int h,int bpp,unsigned int fmt){ int y; for(y=0;y<h;y++){ int x; switch(bpp){ case 1: for(x=0;x<w;x++) dst[x]=src[w-x-1]; break; case 2: switch(fmt){ case IMGFMT_UYVY: { // packed YUV is tricky. U,V are 32bpp while Y is 16bpp: int w2=w>>1; for(x=0;x<w2;x++){ // TODO: optimize this... dst[x*4+0]=src[0+(w2-x-1)*4]; dst[x*4+1]=src[3+(w2-x-1)*4]; dst[x*4+2]=src[2+(w2-x-1)*4]; dst[x*4+3]=src[1+(w2-x-1)*4]; } break; } case IMGFMT_YUY2: case IMGFMT_YVYU: { // packed YUV is tricky. U,V are 32bpp while Y is 16bpp: int w2=w>>1; for(x=0;x<w2;x++){ // TODO: optimize this... dst[x*4+0]=src[2+(w2-x-1)*4]; dst[x*4+1]=src[1+(w2-x-1)*4]; dst[x*4+2]=src[0+(w2-x-1)*4]; dst[x*4+3]=src[3+(w2-x-1)*4]; } break; } default: for(x=0;x<w;x++) *((short*)(dst+x*2))=*((short*)(src+(w-x-1)*2)); } break; case 3: for(x=0;x<w;x++){ dst[x*3+0]=src[0+(w-x-1)*3]; dst[x*3+1]=src[1+(w-x-1)*3]; dst[x*3+2]=src[2+(w-x-1)*3]; } break; case 4: for(x=0;x<w;x++) *((int*)(dst+x*4))=*((int*)(src+(w-x-1)*4)); } src+=srcstride; dst+=dststride; } } //===========================================================================// static int put_image(struct vf_instance_s* vf, mp_image_t *mpi, double pts){ mp_image_t *dmpi; // hope we'll get DR buffer: dmpi=vf_get_image(vf->next,mpi->imgfmt, MP_IMGTYPE_TEMP, MP_IMGFLAG_ACCEPT_STRIDE, mpi->w, mpi->h); if(mpi->flags&MP_IMGFLAG_PLANAR){ mirror(dmpi->planes[0],mpi->planes[0], dmpi->stride[0],mpi->stride[0], dmpi->w,dmpi->h,1,mpi->imgfmt); mirror(dmpi->planes[1],mpi->planes[1], dmpi->stride[1],mpi->stride[1], dmpi->w>>mpi->chroma_x_shift,dmpi->h>>mpi->chroma_y_shift,1,mpi->imgfmt); mirror(dmpi->planes[2],mpi->planes[2], dmpi->stride[2],mpi->stride[2], dmpi->w>>mpi->chroma_x_shift,dmpi->h>>mpi->chroma_y_shift,1,mpi->imgfmt); } else { mirror(dmpi->planes[0],mpi->planes[0], dmpi->stride[0],mpi->stride[0], dmpi->w,dmpi->h,dmpi->bpp>>3,mpi->imgfmt); dmpi->planes[1]=mpi->planes[1]; // passthrough rgb8 palette } return vf_next_put_image(vf,dmpi, pts); } //===========================================================================// static int open(vf_instance_t *vf, char* args){ //vf->config=config; vf->put_image=put_image; return 1; } const vf_info_t vf_info_mirror = { "horizontal mirror", "mirror", "Eyck", "", open, NULL }; //===========================================================================//