#endif
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
+ asm volatile(PREFETCH" %0"::"m"(*s):"memory");
mm_end = end - 23;
- __asm __volatile("movq %0, %%mm7"::"m"(mask32):"memory");
+ asm volatile("movq %0, %%mm7"::"m"(mask32):"memory");
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"punpckldq 3%1, %%mm0 \n\t"
dest += 32;
s += 24;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
#endif
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
+ asm volatile(PREFETCH" %0"::"m"(*s):"memory");
mm_end = end - 31;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq 8%1, %%mm1 \n\t"
dest += 24;
s += 32;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
const uint8_t *mm_end;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s));
- __asm __volatile("movq %0, %%mm4"::"m"(mask15s));
+ asm volatile(PREFETCH" %0"::"m"(*s));
+ asm volatile("movq %0, %%mm4"::"m"(mask15s));
mm_end = end - 15;
while (s<mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq 8%1, %%mm2 \n\t"
d+=16;
s+=16;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
mm_end = end - 3;
while (s < mm_end)
const uint8_t *mm_end;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s));
- __asm __volatile("movq %0, %%mm7"::"m"(mask15rg));
- __asm __volatile("movq %0, %%mm6"::"m"(mask15b));
+ asm volatile(PREFETCH" %0"::"m"(*s));
+ asm volatile("movq %0, %%mm7"::"m"(mask15rg));
+ asm volatile("movq %0, %%mm6"::"m"(mask15b));
mm_end = end - 15;
while (s<mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq 8%1, %%mm2 \n\t"
d+=16;
s+=16;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
mm_end = end - 3;
while (s < mm_end)
: "r" (mm_end), "m" (mask3216g), "m" (mask3216br), "m" (mul3216)
);
#else
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_16mask),"m"(green_16mask));
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 4%1, %%mm3 \n\t"
s += 16;
}
#endif
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
uint16_t *d = (uint16_t *)dst;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_16mask),"m"(green_16mask));
mm_end = end - 15;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 4%1, %%mm3 \n\t"
d += 4;
s += 16;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
: "r" (mm_end), "m" (mask3215g), "m" (mask3216br), "m" (mul3215)
);
#else
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_15mask),"m"(green_15mask));
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 4%1, %%mm3 \n\t"
s += 16;
}
#endif
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
uint16_t *d = (uint16_t *)dst;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_15mask),"m"(green_15mask));
mm_end = end - 15;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 4%1, %%mm3 \n\t"
d += 4;
s += 16;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
uint16_t *d = (uint16_t *)dst;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_16mask),"m"(green_16mask));
mm_end = end - 11;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 3%1, %%mm3 \n\t"
d += 4;
s += 12;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
uint16_t *d = (uint16_t *)dst;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_16mask),"m"(green_16mask));
mm_end = end - 15;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 3%1, %%mm3 \n\t"
d += 4;
s += 12;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
uint16_t *d = (uint16_t *)dst;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_15mask),"m"(green_15mask));
mm_end = end - 11;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 3%1, %%mm3 \n\t"
d += 4;
s += 12;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
uint16_t *d = (uint16_t *)dst;
end = s + src_size;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
+ asm volatile(PREFETCH" %0"::"m"(*src):"memory");
+ asm volatile(
"movq %0, %%mm7 \n\t"
"movq %1, %%mm6 \n\t"
::"m"(red_15mask),"m"(green_15mask));
mm_end = end - 15;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movd %1, %%mm0 \n\t"
"movd 3%1, %%mm3 \n\t"
d += 4;
s += 12;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
const uint16_t *s = (uint16_t *)src;
end = s + src_size/2;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
+ asm volatile(PREFETCH" %0"::"m"(*s):"memory");
mm_end = end - 7;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq %1, %%mm1 \n\t"
:"m"(*s),"m"(mask15b),"m"(mask15g),"m"(mask15r), "m"(mmx_null)
:"memory");
/* Borrowed 32 to 24 */
- __asm __volatile(
+ asm volatile(
"movq %%mm0, %%mm4 \n\t"
"movq %%mm3, %%mm5 \n\t"
"movq %%mm6, %%mm0 \n\t"
d += 24;
s += 8;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
const uint16_t *s = (const uint16_t *)src;
end = s + src_size/2;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
+ asm volatile(PREFETCH" %0"::"m"(*s):"memory");
mm_end = end - 7;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq %1, %%mm1 \n\t"
:"m"(*s),"m"(mask16b),"m"(mask16g),"m"(mask16r),"m"(mmx_null)
:"memory");
/* Borrowed 32 to 24 */
- __asm __volatile(
+ asm volatile(
"movq %%mm0, %%mm4 \n\t"
"movq %%mm3, %%mm5 \n\t"
"movq %%mm6, %%mm0 \n\t"
d += 24;
s += 8;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
const uint16_t *s = (const uint16_t *)src;
end = s + src_size/2;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
- __asm __volatile("pxor %%mm7,%%mm7 \n\t":::"memory");
+ asm volatile(PREFETCH" %0"::"m"(*s):"memory");
+ asm volatile("pxor %%mm7,%%mm7 \n\t":::"memory");
mm_end = end - 3;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq %1, %%mm1 \n\t"
d += 16;
s += 4;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
const uint16_t *s = (uint16_t *)src;
end = s + src_size/2;
#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
- __asm __volatile("pxor %%mm7,%%mm7 \n\t":::"memory");
+ asm volatile(PREFETCH" %0"::"m"(*s):"memory");
+ asm volatile("pxor %%mm7,%%mm7 \n\t":::"memory");
mm_end = end - 3;
while (s < mm_end)
{
- __asm __volatile(
+ asm volatile(
PREFETCH" 32%1 \n\t"
"movq %1, %%mm0 \n\t"
"movq %1, %%mm1 \n\t"
d += 16;
s += 4;
}
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
#endif
while (s < end)
{
long idx = 15 - src_size;
uint8_t *s = (uint8_t *) src-idx, *d = dst-idx;
#ifdef HAVE_MMX
- __asm __volatile(
+ asm volatile(
"test %0, %0 \n\t"
"jns 2f \n\t"
PREFETCH" (%1, %0) \n\t"
: "r" (src-mmx_size), "r"(dst-mmx_size)
);
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory");
+ asm volatile(SFENCE:::"memory");
+ asm volatile(EMMS:::"memory");
if (mmx_size==23) return; //finihsed, was multiple of 8