****************************************************************************/
/* vertical filter [-1 -2 96 42 -7 0] */
-#define QPEL_CAVSV1(A,B,C,D,E,F,OP) \
+#define QPEL_CAVSV1(A,B,C,D,E,F,OP,MUL2) \
"movd (%0), "#F" \n\t"\
"movq "#C", %%mm6 \n\t"\
"pmullw %5, %%mm6 \n\t"\
"movq "#D", %%mm7 \n\t"\
- "pmullw %6, %%mm7 \n\t"\
+ "pmullw "MANGLE(MUL2)", %%mm7\n\t"\
"psllw $3, "#E" \n\t"\
"psubw "#E", %%mm6 \n\t"\
"psraw $3, "#E" \n\t"\
"add %3, %1 \n\t"
/* vertical filter [ 0 -1 5 5 -1 0] */
-#define QPEL_CAVSV2(A,B,C,D,E,F,OP) \
+#define QPEL_CAVSV2(A,B,C,D,E,F,OP,MUL2) \
"movd (%0), "#F" \n\t"\
"movq "#C", %%mm6 \n\t"\
"paddw "#D", %%mm6 \n\t"\
"add %3, %1 \n\t"
/* vertical filter [ 0 -7 42 96 -2 -1] */
-#define QPEL_CAVSV3(A,B,C,D,E,F,OP) \
+#define QPEL_CAVSV3(A,B,C,D,E,F,OP,MUL2) \
"movd (%0), "#F" \n\t"\
"movq "#C", %%mm6 \n\t"\
- "pmullw %6, %%mm6 \n\t"\
+ "pmullw "MANGLE(MUL2)", %%mm6\n\t"\
"movq "#D", %%mm7 \n\t"\
"pmullw %5, %%mm7 \n\t"\
"psllw $3, "#B" \n\t"\
"punpcklbw %%mm7, %%mm2 \n\t"\
"punpcklbw %%mm7, %%mm3 \n\t"\
"punpcklbw %%mm7, %%mm4 \n\t"\
- VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP)\
- VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP)\
- VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP)\
- VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP)\
- VOP(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP)\
- VOP(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP)\
- VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP)\
- VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP)\
+ VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP, MUL2)\
+ VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP, MUL2)\
+ VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP, MUL2)\
+ VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP, MUL2)\
+ VOP(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP, MUL2)\
+ VOP(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP, MUL2)\
+ VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP, MUL2)\
+ VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP, MUL2)\
\
: "+a"(src), "+c"(dst)\
- : "S"((x86_reg)srcStride), "r"((x86_reg)dstStride), "m"(ADD), "m"(MUL1), "m"(MUL2)\
+ : "S"((x86_reg)srcStride), "r"((x86_reg)dstStride), "m"(ADD), "m"(MUL1)\
: "memory"\
);\
if(h==16){\
__asm__ volatile(\
- VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP)\
- VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP)\
- VOP(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP)\
- VOP(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP)\
- VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP)\
- VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP)\
- VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP)\
- VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP)\
+ VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP, MUL2)\
+ VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP, MUL2)\
+ VOP(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP, MUL2)\
+ VOP(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP, MUL2)\
+ VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP, MUL2)\
+ VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP, MUL2)\
+ VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP, MUL2)\
+ VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP, MUL2)\
\
: "+a"(src), "+c"(dst)\
- : "S"((x86_reg)srcStride), "r"((x86_reg)dstStride), "m"(ADD), "m"(MUL1), "m"(MUL2)\
+ : "S"((x86_reg)srcStride), "r"((x86_reg)dstStride), "m"(ADD), "m"(MUL1)\
: "memory"\
);\
}\