vs2010,x64，把内嵌汇编的汇编单独放在.asm文件中求助

void yuv2rgb_32(uint8_t *puc_y, int stride_y, 

				uint8_t *puc_u, uint8_t *puc_v, int stride_uv, 

				uint8_t *puc_out, int width_y, int height_y,

								unsigned int _stride_out) {



	int y, horiz_count;

	int stride_out = width_y * 4;



	if (height_y < 0) {

		/* we are flipping our output upside-down */

		height_y  = -height_y;

		puc_y     += (height_y   - 1) * stride_y ;

		puc_u     += (height_y/2 - 1) * stride_uv;

		puc_v     += (height_y/2 - 1) * stride_uv;

		stride_y  = -stride_y;

		stride_uv = -stride_uv;

	}



	horiz_count = -(width_y >> 3);



	for (y=0; y<height_y; y++) {

	

		_asm {

			push eax

			push ebx

			push ecx

			push edx

			push edi



			mov eax, puc_out       

			mov ebx, puc_y       

			mov ecx, puc_u       

			mov edx, puc_v

			mov edi, horiz_count

			

		horiz_loop:



			movd mm2, [ecx]

			pxor mm7, mm7



			movd mm3, [edx]

			punpcklbw mm2, mm7       ; mm2 = __u3__u2__u1__u0



			movq mm0, [ebx]          ; mm0 = y7y6y5y4y3y2y1y0  

			punpcklbw mm3, mm7       ; mm3 = __v3__v2__v1__v0



			movq mm1, mmw_0x00ff     ; mm1 = 00ff00ff00ff00ff 



			psubusb mm0, mmb_0x10    ; mm0 -= 16



			psubw mm2, mmw_0x0080    ; mm2 -= 128

			pand mm1, mm0            ; mm1 = __y6__y4__y2__y0



			psubw mm3, mmw_0x0080    ; mm3 -= 128

			psllw mm1, 3             ; mm1 *= 8



			psrlw mm0, 8             ; mm0 = __y7__y5__y3__y1

			psllw mm2, 3             ; mm2 *= 8



			pmulhw mm1, mmw_mult_Y   ; mm1 *= luma coeff 

			psllw mm0, 3             ; mm0 *= 8



			psllw mm3, 3             ; mm3 *= 8

			movq mm5, mm3            ; mm5 = mm3 = v



			pmulhw mm5, mmw_mult_V_R ; mm5 = red chroma

			movq mm4, mm2            ; mm4 = mm2 = u



			pmulhw mm0, mmw_mult_Y   ; mm0 *= luma coeff 

			movq mm7, mm1            ; even luma part



			pmulhw mm2, mmw_mult_U_G ; mm2 *= u green coeff 

			paddsw mm7, mm5          ; mm7 = luma + chroma    __r6__r4__r2__r0



			pmulhw mm3, mmw_mult_V_G ; mm3 *= v green coeff  

			packuswb mm7, mm7        ; mm7 = r6r4r2r0r6r4r2r0



			pmulhw mm4, mmw_mult_U_B ; mm4 = blue chroma

			paddsw mm5, mm0          ; mm5 = luma + chroma    __r7__r5__r3__r1



			packuswb mm5, mm5        ; mm6 = r7r5r3r1r7r5r3r1

			paddsw mm2, mm3          ; mm2 = green chroma



			movq mm3, mm1            ; mm3 = __y6__y4__y2__y0

			movq mm6, mm1            ; mm6 = __y6__y4__y2__y0



			paddsw mm3, mm4          ; mm3 = luma + chroma    __b6__b4__b2__b0

			paddsw mm6, mm2          ; mm6 = luma + chroma    __g6__g4__g2__g0

			

			punpcklbw mm7, mm5       ; mm7 = r7r6r5r4r3r2r1r0

			paddsw mm2, mm0          ; odd luma part plus chroma part    __g7__g5__g3__g1



			packuswb mm6, mm6        ; mm2 = g6g4g2g0g6g4g2g0

			packuswb mm2, mm2        ; mm2 = g7g5g3g1g7g5g3g1



			packuswb mm3, mm3        ; mm3 = b6b4b2b0b6b4b2b0

			paddsw mm4, mm0          ; odd luma part plus chroma part    __b7__b5__b3__b1



			packuswb mm4, mm4        ; mm4 = b7b5b3b1b7b5b3b1

			punpcklbw mm6, mm2       ; mm6 = g7g6g5g4g3g2g1g0



			punpcklbw mm3, mm4       ; mm3 = b7b6b5b4b3b2b1b0



			/* 32-bit shuffle.... */

			pxor mm0, mm0            ; is this needed?



			movq mm1, mm6            ; mm1 = g7g6g5g4g3g2g1g0

			punpcklbw mm1, mm0       ; mm1 = __g3__g2__g1__g0



			movq mm0, mm3            ; mm0 = b7b6b5b4b3b2b1b0

			punpcklbw mm0, mm7       ; mm0 = r3b3r2b2r1b1r0b0



			movq mm2, mm0            ; mm2 = r3b3r2b2r1b1r0b0



			punpcklbw mm0, mm1       ; mm0 = __r1g1b1__r0g0b0

			punpckhbw mm2, mm1       ; mm2 = __r3g3b3__r2g2b2



			/* 32-bit save... */

			movq  [eax], mm0         ; eax[0] = __r1g1b1__r0g0b0

			movq mm1, mm6            ; mm1 = g7g6g5g4g3g2g1g0



			movq 8[eax], mm2         ; eax[8] = __r3g3b3__r2g2b2



			/* 32-bit shuffle.... */

			pxor mm0, mm0            ; is this needed?



			punpckhbw mm1, mm0       ; mm1 = __g7__g6__g5__g4



			movq mm0, mm3            ; mm0 = b7b6b5b4b3b2b1b0

			punpckhbw mm0, mm7       ; mm0 = r7b7r6b6r5b5r4b4



			movq mm2, mm0            ; mm2 = r7b7r6b6r5b5r4b4



			punpcklbw mm0, mm1       ; mm0 = __r5g5b5__r4g4b4

			punpckhbw mm2, mm1       ; mm2 = __r7g7b7__r6g6b6



			/* 32-bit save... */

			add ebx, 8               ; puc_y   += 8;

			add ecx, 4               ; puc_u   += 4;



			movq 16[eax], mm0        ; eax[16] = __r5g5b5__r4g4b4

			add edx, 4               ; puc_v   += 4;



			movq 24[eax], mm2        ; eax[24] = __r7g7b7__r6g6b6

			

			// 0 1 2 3 4 5 6 7 rgb save order



			add eax, 32              ; puc_out += 32



			inc edi

			jne horiz_loop			



			pop edi 

			pop edx 

			pop ecx

			pop ebx 

			pop eax



			emms

						

		}

		puc_y   += stride_y;

		if (y%2) {

			puc_u   += stride_uv;

			puc_v   += stride_uv;

		}

		puc_out += stride_out;

	}

}