dsputil.c

君正早期ucos系统(只有早期的才不没有打包成库),MPLAYER,文件系统,图片解码,浏览,电子书,录音,想学ucos,识货的人就下吧 russblock fmradio explore set

      S32ALN(xr1, xr2, xr1, src_rs1);
      S32ALN(xr4, xr5, xr4, src_rs2);
      Q8AVGR(xr1, xr4, xr1);

      S32ALN(xr2, xr3, xr2, src_rs1);
      S32ALN(xr5, xr6, xr5, src_rs2);
      Q8AVGR(xr2, xr5, xr2);
      
      S32SDIV(xr1, dst, dst_stride, 0);
      S32STD(xr2, dst, 4);
    }
}

static void put_no_rnd_pixels8_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				      int dst_stride, int src_stride1, int src_stride2, int h){
    int i;
    uint32_t src_aln1, src_rs1, src_aln2, src_rs2;

    src_aln1 = (uint32_t)src1 & 0xfffffffc;
    src_rs1 = 4 - ((uint32_t)src1 & 3);
    src_aln2 = (uint32_t)src2 & 0xfffffffc;
    src_rs2 = 4 - ((uint32_t)src2 & 3);
    
    dst -= dst_stride;
    src_aln1 -= src_stride1;
    src_aln2 -= src_stride2;

    for(i=0; i<h; i++) {    
      S32LDIV(xr1, src_aln1, src_stride1, 0);
      S32LDD(xr2, src_aln1, 4);
      S32LDD(xr3, src_aln1, 8);
      S32LDIV(xr4, src_aln2, src_stride2, 0);
      S32LDD(xr5, src_aln2, 4);
      S32LDD(xr6, src_aln2, 8);

      S32ALN(xr1, xr2, xr1, src_rs1);
      S32ALN(xr4, xr5, xr4, src_rs2);
      Q8AVG(xr1, xr4, xr1);

      S32ALN(xr2, xr3, xr2, src_rs1);
      S32ALN(xr5, xr6, xr5, src_rs2);
      Q8AVG(xr2, xr5, xr2);
      
      S32SDIV(xr1, dst, dst_stride, 0);
      S32STD(xr2, dst, 4);
    }
}

static void avg_pixels8_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				      int dst_stride, int src_stride1, int src_stride2, int h){
    int i;
    uint32_t src_aln1, src_rs1, src_aln2, src_rs2;

    src_aln1 = (uint32_t)src1 & 0xfffffffc;
    src_rs1 = 4 - ((uint32_t)src1 & 3);
    src_aln2 = (uint32_t)src2 & 0xfffffffc;
    src_rs2 = 4 - ((uint32_t)src2 & 3);
    
    dst -= dst_stride;
    src_aln1 -= src_stride1;
    src_aln2 -= src_stride2;

    for(i=0; i<h; i++) {    
      S32LDIV(xr1, src_aln1, src_stride1, 0);
      S32LDD(xr2, src_aln1, 4);
      S32LDD(xr3, src_aln1, 8);
      S32LDIV(xr4, src_aln2, src_stride2, 0);
      S32LDD(xr5, src_aln2, 4);
      S32LDD(xr6, src_aln2, 8);
      S32LDIV(xr7, dst, dst_stride, 0);
      S32LDD(xr8, dst, 4);

      S32ALN(xr1, xr2, xr1, src_rs1);
      S32ALN(xr4, xr5, xr4, src_rs2);
      Q8AVGR(xr1, xr4, xr1);
      Q8AVGR(xr1, xr7, xr1);

      S32ALN(xr2, xr3, xr2, src_rs1);
      S32ALN(xr5, xr6, xr5, src_rs2);
      Q8AVGR(xr2, xr5, xr2);
      Q8AVGR(xr2, xr8, xr2);
      
      S32STD(xr1, dst, 0);
      S32STD(xr2, dst, 4);
    }
}

static void avg_no_rnd_pixels8_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				      int dst_stride, int src_stride1, int src_stride2, int h){
    int i;
    uint32_t src_aln1, src_rs1, src_aln2, src_rs2;

    src_aln1 = (uint32_t)src1 & 0xfffffffc;
    src_rs1 = 4 - ((uint32_t)src1 & 3);
    src_aln2 = (uint32_t)src2 & 0xfffffffc;
    src_rs2 = 4 - ((uint32_t)src2 & 3);
    
    dst -= dst_stride;
    src_aln1 -= src_stride1;
    src_aln2 -= src_stride2;

    for(i=0; i<h; i++) {    
      S32LDIV(xr1, src_aln1, src_stride1, 0);
      S32LDD(xr2, src_aln1, 4);
      S32LDD(xr3, src_aln1, 8);
      S32LDIV(xr4, src_aln2, src_stride2, 0);
      S32LDD(xr5, src_aln2, 4);
      S32LDD(xr6, src_aln2, 8);
      S32LDIV(xr7, dst, dst_stride, 0);
      S32LDD(xr8, dst, 4);

      S32ALN(xr1, xr2, xr1, src_rs1);
      S32ALN(xr4, xr5, xr4, src_rs2);
      Q8AVG(xr1, xr4, xr1);
      Q8AVGR(xr1, xr7, xr1);

      S32ALN(xr2, xr3, xr2, src_rs1);
      S32ALN(xr5, xr6, xr5, src_rs2);
      Q8AVG(xr2, xr5, xr2);
      Q8AVGR(xr2, xr8, xr2);
      
      S32STD(xr1, dst, 0);
      S32STD(xr2, dst, 4);
    }
}

static void put_pixels8_xy2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
        int i;
        uint32_t a= AV_RN32(pixels+0);
        uint32_t b= AV_RN32(pixels+1);
        uint32_t c= AV_RN32(pixels+4);
        uint32_t d= AV_RN32(pixels+5);
        uint32_t l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x02020202UL;
        uint32_t h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
        uint32_t l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x02020202UL;
        uint32_t h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);

        uint32_t l0_1 ,h0_1, l1_1, h1_1;
        pixels+=line_size;
        for(i=0; i<h; i+=2){
            uint32_t a= AV_RN32(pixels+0);
            uint32_t b= AV_RN32(pixels+1);
            uint32_t c= AV_RN32(pixels+4);
            uint32_t d= AV_RN32(pixels+5);
            l0_1=  (a&0x03030303UL) + (b&0x03030303UL);
            h0_1= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
            l1_1=  (c&0x03030303UL) + (d&0x03030303UL);
            h1_1= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL);
	    *((uint32_t*)(block+4)) = h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL);
            pixels+=line_size;
            block +=line_size;
            a= AV_RN32(pixels+0);
            b= AV_RN32(pixels+1);
            c= AV_RN32(pixels+4);
            d= AV_RN32(pixels+5);
            l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x02020202UL;
            h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
	    l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x02020202UL;
	    h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL);
	    *((uint32_t*)(block+4)) = h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL);
            pixels+=line_size;
            block +=line_size;
        }
}

static void avg_pixels8_xy2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
        int i;
        uint32_t a= AV_RN32(pixels+0);
        uint32_t b= AV_RN32(pixels+1);
        uint32_t c= AV_RN32(pixels+4);
        uint32_t d= AV_RN32(pixels+5);
        uint32_t l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x02020202UL;
        uint32_t h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
        uint32_t l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x02020202UL;
        uint32_t h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);

        uint32_t l0_1 ,h0_1, l1_1, h1_1;
        pixels+=line_size;
        for(i=0; i<h; i+=2){
            uint32_t a= AV_RN32(pixels+0);
            uint32_t b= AV_RN32(pixels+1);
            uint32_t c= AV_RN32(pixels+4);
            uint32_t d= AV_RN32(pixels+5);
            l0_1=  (a&0x03030303UL) + (b&0x03030303UL);
            h0_1= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
            l1_1=  (c&0x03030303UL) + (d&0x03030303UL);
            h1_1= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = rnd_avg32(*((uint32_t*)block), 
					  h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)(block+4)) = rnd_avg32(*((uint32_t*)(block+4)), 
					      h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL));
            pixels+=line_size;
            block +=line_size;
            a= AV_RN32(pixels+0);
            b= AV_RN32(pixels+1);
            c= AV_RN32(pixels+4);
            d= AV_RN32(pixels+5);
            l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x02020202UL;
            h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
	    l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x02020202UL;
	    h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = rnd_avg32(*((uint32_t*)block), 
					  h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)(block+4)) = rnd_avg32(*((uint32_t*)(block+4)), 
					      h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL));
            pixels+=line_size;
            block +=line_size;
        }
}

static void put_no_rnd_pixels8_xy2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
        int i;
        uint32_t a= AV_RN32(pixels+0);
        uint32_t b= AV_RN32(pixels+1);
        uint32_t c= AV_RN32(pixels+4);
        uint32_t d= AV_RN32(pixels+5);
        uint32_t l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x01010101UL;
        uint32_t h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
        uint32_t l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x01010101UL;
        uint32_t h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);

        uint32_t l0_1 ,h0_1, l1_1, h1_1;
        pixels+=line_size;
        for(i=0; i<h; i+=2){
            uint32_t a= AV_RN32(pixels+0);
            uint32_t b= AV_RN32(pixels+1);
            uint32_t c= AV_RN32(pixels+4);
            uint32_t d= AV_RN32(pixels+5);
            l0_1=  (a&0x03030303UL) + (b&0x03030303UL);
            h0_1= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
            l1_1=  (c&0x03030303UL) + (d&0x03030303UL);
            h1_1= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL);
	    *((uint32_t*)(block+4)) = h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL);
            pixels+=line_size;
            block +=line_size;
            a= AV_RN32(pixels+0);
            b= AV_RN32(pixels+1);
            c= AV_RN32(pixels+4);
            d= AV_RN32(pixels+5);
            l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x01010101UL;
            h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
	    l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x01010101UL;
	    h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL);
	    *((uint32_t*)(block+4)) = h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL);
            pixels+=line_size;
            block +=line_size;
        }
}

static void avg_no_rnd_pixels8_xy2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
        int i;
        uint32_t a= AV_RN32(pixels+0);
        uint32_t b= AV_RN32(pixels+1);
        uint32_t c= AV_RN32(pixels+4);
        uint32_t d= AV_RN32(pixels+5);
        uint32_t l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x01010101UL;
        uint32_t h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
        uint32_t l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x01010101UL;
        uint32_t h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);

        uint32_t l0_1 ,h0_1, l1_1, h1_1;
        pixels+=line_size;
        for(i=0; i<h; i+=2){
            uint32_t a= AV_RN32(pixels+0);
            uint32_t b= AV_RN32(pixels+1);
            uint32_t c= AV_RN32(pixels+4);
            uint32_t d= AV_RN32(pixels+5);
            l0_1=  (a&0x03030303UL) + (b&0x03030303UL);
            h0_1= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
            l1_1=  (c&0x03030303UL) + (d&0x03030303UL);
            h1_1= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = rnd_avg32(*((uint32_t*)block), 
					  h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)(block+4)) = rnd_avg32(*((uint32_t*)(block+4)), 
					      h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL));
            pixels+=line_size;
            block +=line_size;
            a= AV_RN32(pixels+0);
            b= AV_RN32(pixels+1);
            c= AV_RN32(pixels+4);
            d= AV_RN32(pixels+5);
            l0_0=  (a&0x03030303UL) + (b&0x03030303UL) + 0x01010101UL;
            h0_0= ((a&0xFCFCFCFCUL)>>2) + ((b&0xFCFCFCFCUL)>>2);
	    l1_0=  (c&0x03030303UL) + (d&0x03030303UL) + 0x01010101UL;
	    h1_0= ((c&0xFCFCFCFCUL)>>2) + ((d&0xFCFCFCFCUL)>>2);
	    //            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)block) = rnd_avg32(*((uint32_t*)block), 
					  h0_0+h0_1+(((l0_0+l0_1)>>2)&0x0F0F0F0FUL));
	    *((uint32_t*)(block+4)) = rnd_avg32(*((uint32_t*)(block+4)), 
					      h1_0+h1_1+(((l1_0+l1_1)>>2)&0x0F0F0F0FUL));
            pixels+=line_size;
            block +=line_size;
        }
}

static void put_pixels16_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				int dst_stride, int src_stride1, int src_stride2, int h){
    put_pixels8_l2_mxu(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);
    put_pixels8_l2_mxu(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);
}
static void avg_pixels16_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				int dst_stride, int src_stride1, int src_stride2, int h){
    avg_pixels8_l2_mxu(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);
    avg_pixels8_l2_mxu(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);
}
static void put_no_rnd_pixels16_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				int dst_stride, int src_stride1, int src_stride2, int h){
    put_no_rnd_pixels8_l2_mxu(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);
    put_no_rnd_pixels8_l2_mxu(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);
}
static void avg_no_rnd_pixels16_l2_mxu(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, 
				int dst_stride, int src_stride1, int src_stride2, int h){
    avg_no_rnd_pixels8_l2_mxu(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);
    avg_no_rnd_pixels8_l2_mxu(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);
}

static void put_pixels4_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    put_pixels4_l2_mxu(block, pixels, pixels+1, line_size, line_size, line_size, h);
}

static void put_pixels4_y2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    put_pixels4_l2_mxu(block, pixels, pixels+line_size, line_size, line_size, line_size, h);
}

static void put_pixels2_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    put_pixels2_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);
}

static void put_pixels2_y2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    put_pixels2_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);
}

static void avg_pixels4_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    avg_pixels4_l2_mxu(block, pixels, pixels+1, line_size, line_size, line_size, h);
}

static void avg_pixels4_y2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    avg_pixels4_l2_mxu(block, pixels, pixels+line_size, line_size, line_size, line_size, h);
}

static void avg_pixels2_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    avg_pixels2_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);
}

static void avg_pixels2_y2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    avg_pixels2_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);
}

static void put_no_rnd_pixels8_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    put_no_rnd_pixels8_l2_mxu(block, pixels, pixels+1, line_size, line_size, line_size, h);
}
static void avg_no_rnd_pixels8_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    avg_no_rnd_pixels8_l2_mxu(block, pixels, pixels+1, line_size, line_size, line_size, h);
}
static void put_pixels8_x2_mxu(uint8_t *block, const uint8_t *pixels, int line_size, int h){
    put_pixels8_l2_mxu(block, pixels, pixels+1, line_size, line_size, line_size, h);
}
static void av