mirror of
https://github.com/DrBeef/Raze.git
synced 2024-12-15 15:11:01 +00:00
98ebda94ec
git-svn-id: https://svn.eduke32.com/eduke32@5551 1a8010ca-5511-0410-912e-c29ae57300e0
90 lines
3 KiB
C
90 lines
3 KiB
C
//
|
|
// GCC Inline Assembler version (ARM)
|
|
//
|
|
|
|
#ifndef pragmas_arm_h_
|
|
#define pragmas_arm_h_
|
|
|
|
#define EDUKE32_SCALER_PRAGMA(a) \
|
|
static inline int32_t mulscale##a(int32_t eax, int32_t edx) { return dw((qw(eax) * edx) >> by(a)); } \
|
|
\
|
|
static inline int32_t dmulscale##a(int32_t eax, int32_t edx, int32_t esi, int32_t edi) \
|
|
{ \
|
|
return dw(((qw(eax) * edx) + (qw(esi) * edi)) >> by(a)); \
|
|
}
|
|
|
|
EDUKE32_GENERATE_PRAGMAS EDUKE32_SCALER_PRAGMA(32)
|
|
|
|
#undef EDUKE32_SCALER_PRAGMA
|
|
|
|
static inline void swapchar(void *a, void *b)
|
|
{
|
|
char t = *((char *)b);
|
|
*((char *)b) = *((char *)a);
|
|
*((char *)a) = t;
|
|
}
|
|
static inline void swapchar2(void *a, void *b, int32_t s)
|
|
{
|
|
swapchar(a, b);
|
|
swapchar((char *)a + 1, (char *)b + s);
|
|
}
|
|
static inline void swapshort(void *a, void *b)
|
|
{
|
|
int16_t t = *((int16_t *)b);
|
|
*((int16_t *)b) = *((int16_t *)a);
|
|
*((int16_t *)a) = t;
|
|
}
|
|
static inline void swaplong(void *a, void *b)
|
|
{
|
|
int32_t t = *((int32_t *)b);
|
|
*((int32_t *)b) = *((int32_t *)a);
|
|
*((int32_t *)a) = t;
|
|
}
|
|
static inline void swapfloat(void *a, void *b)
|
|
{
|
|
float t = *((float *)b);
|
|
*((float *)b) = *((float *)a);
|
|
*((float *)a) = t;
|
|
}
|
|
static inline void swap64bit(void *a, void *b)
|
|
{
|
|
int64_t t = *((int64_t *)b);
|
|
*((int64_t *)b) = *((int64_t *)a);
|
|
*((int64_t *)a) = t;
|
|
}
|
|
|
|
static inline char readpixel(void *s) { return (*((char *)(s))); }
|
|
static inline void drawpixel(void *s, char a) { *((char *)(s)) = a; }
|
|
|
|
static inline int32_t klabs(int32_t a)
|
|
{
|
|
const uint32_t m = a >> (sizeof(int32_t) * CHAR_BIT - 1);
|
|
return (a ^ m) - m;
|
|
}
|
|
static inline int32_t ksgn(int32_t a) { return (a > 0) - (a < 0); }
|
|
|
|
static inline int32_t mulscale(int32_t eax, int32_t edx, int32_t ecx) { return dw((qw(eax) * qw(edx)) >> by(ecx)); }
|
|
static inline int32_t dmulscale(int32_t eax, int32_t edx, int32_t esi, int32_t edi, int32_t ecx)
|
|
{
|
|
return dw(((qw(eax) * qw(edx)) + (qw(esi) * qw(edi))) >> by(ecx));
|
|
}
|
|
|
|
void qinterpolatedown16(intptr_t bufptr, int32_t num, int32_t val, int32_t add);
|
|
void qinterpolatedown16short(intptr_t bufptr, int32_t num, int32_t val, int32_t add);
|
|
|
|
void clearbuf(void *d, int32_t c, int32_t a);
|
|
void copybuf(const void *s, void *d, int32_t c);
|
|
void swapbuf4(void *a, void *b, int32_t c);
|
|
|
|
void clearbufbyte(void *D, int32_t c, int32_t a);
|
|
void copybufbyte(const void *S, void *D, int32_t c);
|
|
void copybufreverse(const void *S, void *D, int32_t c);
|
|
|
|
static inline int32_t krecipasm(int32_t i)
|
|
{
|
|
// Ken did this
|
|
float const f = (float const)i;
|
|
i = *(int32_t const *)&f;
|
|
return ((reciptable[(i >> 12) & 2047] >> (((i - 0x3f800000) >> 23) & 31)) ^ (i >> 31));
|
|
}
|
|
#endif
|