#include <config.h>
#endif
-#include <pulse/rtclock.h>
-
#include <pulsecore/macro.h>
#include <pulsecore/endianmacros.h>
"movs %[n], %[n], lsr #2 \n\t"
"beq 2f \n\t"
- "vdup.f32 q1, %[scale] \n\t"
-
"1: \n\t"
"vld1.32 {q0}, [%[src]]! \n\t"
- "vmul.f32 q0, q0, q1 \n\t" /* scale */
- "vcvt.s32.f32 q0, q0, #16 \n\t" /* s32<-f32 as 16:16 fixed-point */
+ "vcvt.s32.f32 q0, q0, #31 \n\t" /* s32<-f32 as 16:16 fixed-point, with implicit multiplication by 32768 */
"vqrshrn.s32 d0, q0, #16 \n\t" /* shift, round, narrow */
"subs %[n], %[n], #1 \n\t"
"vst1.16 {d0}, [%[dst]]! \n\t"
"2: \n\t"
: [dst] "+r" (dst), [src] "+r" (src), [n] "+r" (n) /* output operands (or input operands that get modified) */
- : [scale] "r" (32768.0f) /* input operands */
- : "memory", "cc", "q0", "q1" /* clobber list */
+ : /* input operands */
+ : "memory", "cc", "q0" /* clobber list */
);
/* leftovers */
static void pa_sconv_s16le_to_f32ne_neon(unsigned n, const int16_t *src, float *dst) {
unsigned i = n & 3;
-
const float invscale = 1.0f / (1 << 15);
__asm__ __volatile__ (
- "movs %[n], %[n], lsr #2 \n\t"
+ "movs %[n], %[n], lsr #2 \n\t"
"beq 2f \n\t"
- "vdup.f32 q1, %[invscale] \n\t"
-
"1: \n\t"
"vld1.16 {d0}, [%[src]]! \n\t"
"vmovl.s16 q0, d0 \n\t" /* widen */
- "vcvt.f32.s32 q0, q0 \n\t" /* f32<-s32 */
- "vmul.f32 q0, q0, q1 \n\t"
+ "vcvt.f32.s32 q0, q0, #15 \n\t" /* f32<-s32 and divide by (1<<15) */
"subs %[n], %[n], #1 \n\t"
"vst1.32 {q0}, [%[dst]]! \n\t"
"bgt 1b \n\t"
"2: \n\t"
: [dst] "+r" (dst), [src] "+r" (src), [n] "+r" (n) /* output operands (or input operands that get modified) */
- : [invscale] "r" (invscale) /* input operands */
- : "memory", "cc", "q0", "q1" /* clobber list */
+ : /* input operands */
+ : "memory", "cc", "q0" /* clobber list */
);
/* leftovers */