diff --git a/src/stdlib/SDL_stdlib.c b/src/stdlib/SDL_stdlib.c index 22eab5b02..8d9707e78 100644 --- a/src/stdlib/SDL_stdlib.c +++ b/src/stdlib/SDL_stdlib.c @@ -486,42 +486,22 @@ int SDL_tolower(int x) { return ((x) >= 'A') && ((x) <= 'Z') ? ('a'+((x)-'A')) : __declspec(selectany) int _fltused = 1; #endif -/* The optimizer on Visual Studio 2005 and later generates memcpy() calls */ -#if (_MSC_VER >= 1400) && defined(_WIN64) && !defined(_DEBUG) && !(_MSC_VER >= 1900 && defined(_MT)) -#include - +/* The optimizer on Visual Studio 2005 and later generates memcpy() and memset() calls */ +#if _MSC_VER >= 1400 #pragma function(memcpy) -void * memcpy ( void * destination, const void * source, size_t num ) +void * +memcpy(void *dst, const void *src, size_t len) { - const Uint8 *src = (const Uint8 *)source; - Uint8 *dst = (Uint8 *)destination; - size_t i; - - /* All WIN64 architectures have SSE, right? */ - if (!((uintptr_t) src & 15) && !((uintptr_t) dst & 15)) { - __m128 values[4]; - for (i = num / 64; i--;) { - _mm_prefetch(src, _MM_HINT_NTA); - values[0] = *(__m128 *) (src + 0); - values[1] = *(__m128 *) (src + 16); - values[2] = *(__m128 *) (src + 32); - values[3] = *(__m128 *) (src + 48); - _mm_stream_ps((float *) (dst + 0), values[0]); - _mm_stream_ps((float *) (dst + 16), values[1]); - _mm_stream_ps((float *) (dst + 32), values[2]); - _mm_stream_ps((float *) (dst + 48), values[3]); - src += 64; - dst += 64; - } - num &= 63; - } - - while (num--) { - *dst++ = *src++; - } - return destination; + return SDL_memcpy(dst, src, len); } -#endif /* _MSC_VER == 1600 && defined(_WIN64) && !defined(_DEBUG) */ + +#pragma function(memset) +void * +memset(void *dst, int c, size_t len) +{ + return SDL_memset(dst, c, len); +} +#endif #ifdef _M_IX86