return result;
}
-#define vfma_lane_f32(a, b, c, d) \
- __extension__ \
- ({ \
- float32x2_t c_ = (c); \
- float32x2_t b_ = (b); \
- float32x2_t a_ = (a); \
- float32x2_t result; \
- __asm__ ("fmla %0.2s,%2.2s,%3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vfmad_lane_f64(a, b, c) \
- __extension__ \
- ({ \
- float64x2_t b_ = (b); \
- float64_t a_ = (a); \
- float64_t result; \
- __asm__ ("fmla %d0,%d1,%2.d[%3]" \
- : "=w"(result) \
- : "w"(a_), "w"(b_), "i"(c) \
- : /* No clobbers */); \
- result; \
- })
-
__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
vfmaq_f32 (float32x4_t a, float32x4_t b, float32x4_t c)
{
return result;
}
-#define vfmaq_lane_f32(a, b, c, d) \
- __extension__ \
- ({ \
- float32x4_t c_ = (c); \
- float32x4_t b_ = (b); \
- float32x4_t a_ = (a); \
- float32x4_t result; \
- __asm__ ("fmla %0.4s,%2.4s,%3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vfmaq_lane_f64(a, b, c, d) \
- __extension__ \
- ({ \
- float64x2_t c_ = (c); \
- float64x2_t b_ = (b); \
- float64x2_t a_ = (a); \
- float64x2_t result; \
- __asm__ ("fmla %0.2d,%2.2d,%3.d[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vfmas_lane_f32(a, b, c) \
- __extension__ \
- ({ \
- float32x4_t b_ = (b); \
- float32_t a_ = (a); \
- float32_t result; \
- __asm__ ("fmla %s0,%s1,%2.s[%3]" \
- : "=w"(result) \
- : "w"(a_), "w"(b_), "i"(c) \
- : /* No clobbers */); \
- result; \
- })
-
__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
vfma_n_f32 (float32x2_t a, float32x2_t b, float32_t c)
{
return result;
}
-#define vfmsd_lane_f64(a, b, c) \
- __extension__ \
- ({ \
- float64x2_t b_ = (b); \
- float64_t a_ = (a); \
- float64_t result; \
- __asm__ ("fmls %d0,%d1,%2.d[%3]" \
- : "=w"(result) \
- : "w"(a_), "w"(b_), "i"(c) \
- : /* No clobbers */); \
- result; \
- })
-
__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
vfmsq_f32 (float32x4_t a, float32x4_t b, float32x4_t c)
{
return result;
}
-#define vfmss_lane_f32(a, b, c) \
- __extension__ \
- ({ \
- float32x4_t b_ = (b); \
- float32_t a_ = (a); \
- float32_t result; \
- __asm__ ("fmls %s0,%s1,%2.s[%3]" \
- : "=w"(result) \
- : "w"(a_), "w"(b_), "i"(c) \
- : /* No clobbers */); \
- result; \
- })
-
__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
vget_high_f32 (float32x4_t a)
{
result; \
})
-#define vmla_lane_f32(a, b, c, d) \
- __extension__ \
- ({ \
- float32x2_t c_ = (c); \
- float32x2_t b_ = (b); \
- float32x2_t a_ = (a); \
- float32x2_t result; \
- float32x2_t t1; \
- __asm__ ("fmul %1.2s, %3.2s, %4.s[%5]; fadd %0.2s, %0.2s, %1.2s" \
- : "=w"(result), "=w"(t1) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_lane_s16(a, b, c, d) \
- __extension__ \
- ({ \
- int16x4_t c_ = (c); \
- int16x4_t b_ = (b); \
- int16x4_t a_ = (a); \
- int16x4_t result; \
- __asm__ ("mla %0.4h, %2.4h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_lane_s32(a, b, c, d) \
- __extension__ \
- ({ \
- int32x2_t c_ = (c); \
- int32x2_t b_ = (b); \
- int32x2_t a_ = (a); \
- int32x2_t result; \
- __asm__ ("mla %0.2s, %2.2s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_lane_u16(a, b, c, d) \
- __extension__ \
- ({ \
- uint16x4_t c_ = (c); \
- uint16x4_t b_ = (b); \
- uint16x4_t a_ = (a); \
- uint16x4_t result; \
- __asm__ ("mla %0.4h, %2.4h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_lane_u32(a, b, c, d) \
- __extension__ \
- ({ \
- uint32x2_t c_ = (c); \
- uint32x2_t b_ = (b); \
- uint32x2_t a_ = (a); \
- uint32x2_t result; \
- __asm__ ("mla %0.2s, %2.2s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_laneq_s16(a, b, c, d) \
- __extension__ \
- ({ \
- int16x8_t c_ = (c); \
- int16x4_t b_ = (b); \
- int16x4_t a_ = (a); \
- int16x4_t result; \
- __asm__ ("mla %0.4h, %2.4h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_laneq_s32(a, b, c, d) \
- __extension__ \
- ({ \
- int32x4_t c_ = (c); \
- int32x2_t b_ = (b); \
- int32x2_t a_ = (a); \
- int32x2_t result; \
- __asm__ ("mla %0.2s, %2.2s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_laneq_u16(a, b, c, d) \
- __extension__ \
- ({ \
- uint16x8_t c_ = (c); \
- uint16x4_t b_ = (b); \
- uint16x4_t a_ = (a); \
- uint16x4_t result; \
- __asm__ ("mla %0.4h, %2.4h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmla_laneq_u32(a, b, c, d) \
- __extension__ \
- ({ \
- uint32x4_t c_ = (c); \
- uint32x2_t b_ = (b); \
- uint32x2_t a_ = (a); \
- uint32x2_t result; \
- __asm__ ("mla %0.2s, %2.2s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
vmla_n_f32 (float32x2_t a, float32x2_t b, float32_t c)
{
return result;
}
-#define vmlaq_lane_f32(a, b, c, d) \
- __extension__ \
- ({ \
- float32x4_t c_ = (c); \
- float32x4_t b_ = (b); \
- float32x4_t a_ = (a); \
- float32x4_t result; \
- float32x4_t t1; \
- __asm__ ("fmul %1.4s, %3.4s, %4.s[%5]; fadd %0.4s, %0.4s, %1.4s" \
- : "=w"(result), "=w"(t1) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_lane_s16(a, b, c, d) \
- __extension__ \
- ({ \
- int16x8_t c_ = (c); \
- int16x8_t b_ = (b); \
- int16x8_t a_ = (a); \
- int16x8_t result; \
- __asm__ ("mla %0.8h, %2.8h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_lane_s32(a, b, c, d) \
- __extension__ \
- ({ \
- int32x4_t c_ = (c); \
- int32x4_t b_ = (b); \
- int32x4_t a_ = (a); \
- int32x4_t result; \
- __asm__ ("mla %0.4s, %2.4s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_lane_u16(a, b, c, d) \
- __extension__ \
- ({ \
- uint16x8_t c_ = (c); \
- uint16x8_t b_ = (b); \
- uint16x8_t a_ = (a); \
- uint16x8_t result; \
- __asm__ ("mla %0.8h, %2.8h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_lane_u32(a, b, c, d) \
- __extension__ \
- ({ \
- uint32x4_t c_ = (c); \
- uint32x4_t b_ = (b); \
- uint32x4_t a_ = (a); \
- uint32x4_t result; \
- __asm__ ("mla %0.4s, %2.4s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_laneq_s16(a, b, c, d) \
- __extension__ \
- ({ \
- int16x8_t c_ = (c); \
- int16x8_t b_ = (b); \
- int16x8_t a_ = (a); \
- int16x8_t result; \
- __asm__ ("mla %0.8h, %2.8h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_laneq_s32(a, b, c, d) \
- __extension__ \
- ({ \
- int32x4_t c_ = (c); \
- int32x4_t b_ = (b); \
- int32x4_t a_ = (a); \
- int32x4_t result; \
- __asm__ ("mla %0.4s, %2.4s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_laneq_u16(a, b, c, d) \
- __extension__ \
- ({ \
- uint16x8_t c_ = (c); \
- uint16x8_t b_ = (b); \
- uint16x8_t a_ = (a); \
- uint16x8_t result; \
- __asm__ ("mla %0.8h, %2.8h, %3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlaq_laneq_u32(a, b, c, d) \
- __extension__ \
- ({ \
- uint32x4_t c_ = (c); \
- uint32x4_t b_ = (b); \
- uint32x4_t a_ = (a); \
- uint32x4_t result; \
- __asm__ ("mla %0.4s, %2.4s, %3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
vmlaq_n_f32 (float32x4_t a, float32x4_t b, float32_t c)
{
__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
vmlaq_u8 (uint8x16_t a, uint8x16_t b, uint8x16_t c)
{
- uint8x16_t result;
- __asm__ ("mla %0.16b, %2.16b, %3.16b"
- : "=w"(result)
- : "0"(a), "w"(b), "w"(c)
- : /* No clobbers */);
- return result;
-}
-
-__extension__ static __inline uint16x8_t __attribute__ ((__always_inline__))
-vmlaq_u16 (uint16x8_t a, uint16x8_t b, uint16x8_t c)
-{
- uint16x8_t result;
- __asm__ ("mla %0.8h, %2.8h, %3.8h"
- : "=w"(result)
- : "0"(a), "w"(b), "w"(c)
- : /* No clobbers */);
- return result;
-}
-
-__extension__ static __inline uint32x4_t __attribute__ ((__always_inline__))
-vmlaq_u32 (uint32x4_t a, uint32x4_t b, uint32x4_t c)
-{
- uint32x4_t result;
- __asm__ ("mla %0.4s, %2.4s, %3.4s"
- : "=w"(result)
- : "0"(a), "w"(b), "w"(c)
- : /* No clobbers */);
- return result;
-}
-
-#define vmls_lane_f32(a, b, c, d) \
- __extension__ \
- ({ \
- float32x2_t c_ = (c); \
- float32x2_t b_ = (b); \
- float32x2_t a_ = (a); \
- float32x2_t result; \
- float32x2_t t1; \
- __asm__ ("fmul %1.2s, %3.2s, %4.s[%5]; fsub %0.2s, %0.2s, %1.2s" \
- : "=w"(result), "=w"(t1) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmls_lane_s16(a, b, c, d) \
- __extension__ \
- ({ \
- int16x4_t c_ = (c); \
- int16x4_t b_ = (b); \
- int16x4_t a_ = (a); \
- int16x4_t result; \
- __asm__ ("mls %0.4h,%2.4h,%3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmls_lane_s32(a, b, c, d) \
- __extension__ \
- ({ \
- int32x2_t c_ = (c); \
- int32x2_t b_ = (b); \
- int32x2_t a_ = (a); \
- int32x2_t result; \
- __asm__ ("mls %0.2s,%2.2s,%3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmls_lane_u16(a, b, c, d) \
- __extension__ \
- ({ \
- uint16x4_t c_ = (c); \
- uint16x4_t b_ = (b); \
- uint16x4_t a_ = (a); \
- uint16x4_t result; \
- __asm__ ("mls %0.4h,%2.4h,%3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
+ uint8x16_t result;
+ __asm__ ("mla %0.16b, %2.16b, %3.16b"
+ : "=w"(result)
+ : "0"(a), "w"(b), "w"(c)
+ : /* No clobbers */);
+ return result;
+}
-#define vmls_lane_u32(a, b, c, d) \
- __extension__ \
- ({ \
- uint32x2_t c_ = (c); \
- uint32x2_t b_ = (b); \
- uint32x2_t a_ = (a); \
- uint32x2_t result; \
- __asm__ ("mls %0.2s,%2.2s,%3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
+__extension__ static __inline uint16x8_t __attribute__ ((__always_inline__))
+vmlaq_u16 (uint16x8_t a, uint16x8_t b, uint16x8_t c)
+{
+ uint16x8_t result;
+ __asm__ ("mla %0.8h, %2.8h, %3.8h"
+ : "=w"(result)
+ : "0"(a), "w"(b), "w"(c)
+ : /* No clobbers */);
+ return result;
+}
+
+__extension__ static __inline uint32x4_t __attribute__ ((__always_inline__))
+vmlaq_u32 (uint32x4_t a, uint32x4_t b, uint32x4_t c)
+{
+ uint32x4_t result;
+ __asm__ ("mla %0.4s, %2.4s, %3.4s"
+ : "=w"(result)
+ : "0"(a), "w"(b), "w"(c)
+ : /* No clobbers */);
+ return result;
+}
__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
vmls_n_f32 (float32x2_t a, float32x2_t b, float32_t c)
return result;
}
-#define vmlsq_lane_f32(a, b, c, d) \
- __extension__ \
- ({ \
- float32x4_t c_ = (c); \
- float32x4_t b_ = (b); \
- float32x4_t a_ = (a); \
- float32x4_t result; \
- float32x4_t t1; \
- __asm__ ("fmul %1.4s, %3.4s, %4.s[%5]; fsub %0.4s, %0.4s, %1.4s" \
- : "=w"(result), "=w"(t1) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlsq_lane_s16(a, b, c, d) \
- __extension__ \
- ({ \
- int16x8_t c_ = (c); \
- int16x8_t b_ = (b); \
- int16x8_t a_ = (a); \
- int16x8_t result; \
- __asm__ ("mls %0.8h,%2.8h,%3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlsq_lane_s32(a, b, c, d) \
- __extension__ \
- ({ \
- int32x4_t c_ = (c); \
- int32x4_t b_ = (b); \
- int32x4_t a_ = (a); \
- int32x4_t result; \
- __asm__ ("mls %0.4s,%2.4s,%3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlsq_lane_u16(a, b, c, d) \
- __extension__ \
- ({ \
- uint16x8_t c_ = (c); \
- uint16x8_t b_ = (b); \
- uint16x8_t a_ = (a); \
- uint16x8_t result; \
- __asm__ ("mls %0.8h,%2.8h,%3.h[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "x"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlsq_lane_u32(a, b, c, d) \
- __extension__ \
- ({ \
- uint32x4_t c_ = (c); \
- uint32x4_t b_ = (b); \
- uint32x4_t a_ = (a); \
- uint32x4_t result; \
- __asm__ ("mls %0.4s,%2.4s,%3.s[%4]" \
- : "=w"(result) \
- : "0"(a_), "w"(b_), "w"(c_), "i"(d) \
- : /* No clobbers */); \
- result; \
- })
-
-#define vmlsq_laneq_f32(__a, __b, __c, __d) \
- __extension__ \
- ({ \
- float32x4_t __c_ = (__c); \
- float32x4_t __b_ = (__b); \
- float32x4_t __a_ = (__a); \
- float32x4_t __result; \
- float32x4_t __t1; \
- __asm__ ("fmul %1.4s, %3.4s, %4.s[%5]; fsub %0.4s, %0.4s, %1.4s" \
- : "=w"(__result), "=w"(__t1) \
- : "0"(__a_), "w"(__b_), "w"(__c_), "i"(__d) \
- : /* No clobbers */); \
- __result; \
- })
-
-#define vmlsq_laneq_s16(__a, __b, __c, __d) \
- __extension__ \
- ({ \
- int16x8_t __c_ = (__c); \
- int16x8_t __b_ = (__b); \
- int16x8_t __a_ = (__a); \
- int16x8_t __result; \
- __asm__ ("mls %0.8h, %2.8h, %3.h[%4]" \
- : "=w"(__result) \
- : "0"(__a_), "w"(__b_), "x"(__c_), "i"(__d) \
- : /* No clobbers */); \
- __result; \
- })
-
-#define vmlsq_laneq_s32(__a, __b, __c, __d) \
- __extension__ \
- ({ \
- int32x4_t __c_ = (__c); \
- int32x4_t __b_ = (__b); \
- int32x4_t __a_ = (__a); \
- int32x4_t __result; \
- __asm__ ("mls %0.4s, %2.4s, %3.s[%4]" \
- : "=w"(__result) \
- : "0"(__a_), "w"(__b_), "w"(__c_), "i"(__d) \
- : /* No clobbers */); \
- __result; \
- })
-
-#define vmlsq_laneq_u16(__a, __b, __c, __d) \
- __extension__ \
- ({ \
- uint16x8_t __c_ = (__c); \
- uint16x8_t __b_ = (__b); \
- uint16x8_t __a_ = (__a); \
- uint16x8_t __result; \
- __asm__ ("mls %0.8h, %2.8h, %3.h[%4]" \
- : "=w"(__result) \
- : "0"(__a_), "w"(__b_), "x"(__c_), "i"(__d) \
- : /* No clobbers */); \
- __result; \
- })
-
-#define vmlsq_laneq_u32(__a, __b, __c, __d) \
- __extension__ \
- ({ \
- uint32x4_t __c_ = (__c); \
- uint32x4_t __b_ = (__b); \
- uint32x4_t __a_ = (__a); \
- uint32x4_t __result; \
- __asm__ ("mls %0.4s, %2.4s, %3.s[%4]" \
- : "=w"(__result) \
- : "0"(__a_), "w"(__b_), "w"(__c_), "i"(__d) \
- : /* No clobbers */); \
- __result; \
- })
-
__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
vmlsq_n_f32 (float32x4_t a, float32x4_t b, float32_t c)
{
return __aarch64_vget_lane_p16 (__a, __b);
}
-__extension__ static __inline int16_t __attribute__ ((__always_inline__))
-vduph_lane_s16 (int16x4_t __a, const int __b)
+__extension__ static __inline int16_t __attribute__ ((__always_inline__))
+vduph_lane_s16 (int16x4_t __a, const int __b)
+{
+ return __aarch64_vget_lane_s16 (__a, __b);
+}
+
+__extension__ static __inline uint16_t __attribute__ ((__always_inline__))
+vduph_lane_u16 (uint16x4_t __a, const int __b)
+{
+ return __aarch64_vget_lane_u16 (__a, __b);
+}
+
+/* vdups_lane */
+__extension__ static __inline float32_t __attribute__ ((__always_inline__))
+vdups_lane_f32 (float32x2_t __a, const int __b)
+{
+ return __aarch64_vget_lane_f32 (__a, __b);
+}
+
+__extension__ static __inline int32_t __attribute__ ((__always_inline__))
+vdups_lane_s32 (int32x2_t __a, const int __b)
+{
+ return __aarch64_vget_lane_s32 (__a, __b);
+}
+
+__extension__ static __inline uint32_t __attribute__ ((__always_inline__))
+vdups_lane_u32 (uint32x2_t __a, const int __b)
+{
+ return __aarch64_vget_lane_u32 (__a, __b);
+}
+
+/* vdupd_lane */
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vdupd_lane_f64 (float64x1_t __a, const int __attribute__ ((unused)) __b)
+{
+ return __a;
+}
+
+__extension__ static __inline int64_t __attribute__ ((__always_inline__))
+vdupd_lane_s64 (int64x1_t __a, const int __attribute__ ((unused)) __b)
+{
+ return __a;
+}
+
+__extension__ static __inline uint64_t __attribute__ ((__always_inline__))
+vdupd_lane_u64 (uint64x1_t __a, const int __attribute__ ((unused)) __b)
+{
+ return __a;
+}
+
+/* vdupb_laneq */
+__extension__ static __inline poly8_t __attribute__ ((__always_inline__))
+vdupb_laneq_p8 (poly8x16_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_p8 (__a, __b);
+}
+
+__extension__ static __inline int8_t __attribute__ ((__always_inline__))
+vdupb_laneq_s8 (int8x16_t __a, const int __attribute__ ((unused)) __b)
+{
+ return __aarch64_vgetq_lane_s8 (__a, __b);
+}
+
+__extension__ static __inline uint8_t __attribute__ ((__always_inline__))
+vdupb_laneq_u8 (uint8x16_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_u8 (__a, __b);
+}
+
+/* vduph_laneq */
+__extension__ static __inline poly16_t __attribute__ ((__always_inline__))
+vduph_laneq_p16 (poly16x8_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_p16 (__a, __b);
+}
+
+__extension__ static __inline int16_t __attribute__ ((__always_inline__))
+vduph_laneq_s16 (int16x8_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_s16 (__a, __b);
+}
+
+__extension__ static __inline uint16_t __attribute__ ((__always_inline__))
+vduph_laneq_u16 (uint16x8_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_u16 (__a, __b);
+}
+
+/* vdups_laneq */
+__extension__ static __inline float32_t __attribute__ ((__always_inline__))
+vdups_laneq_f32 (float32x4_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_f32 (__a, __b);
+}
+
+__extension__ static __inline int32_t __attribute__ ((__always_inline__))
+vdups_laneq_s32 (int32x4_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_s32 (__a, __b);
+}
+
+__extension__ static __inline uint32_t __attribute__ ((__always_inline__))
+vdups_laneq_u32 (uint32x4_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_u32 (__a, __b);
+}
+
+/* vdupd_laneq */
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vdupd_laneq_f64 (float64x2_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_f64 (__a, __b);
+}
+
+__extension__ static __inline int64_t __attribute__ ((__always_inline__))
+vdupd_laneq_s64 (int64x2_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_s64 (__a, __b);
+}
+
+__extension__ static __inline uint64_t __attribute__ ((__always_inline__))
+vdupd_laneq_u64 (uint64x2_t __a, const int __b)
+{
+ return __aarch64_vgetq_lane_u64 (__a, __b);
+}
+
+/* vfma_lane */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vfma_lane_f32 (float32x2_t __a, float32x2_t __b,
+ float32x2_t __c, const int __lane)
+{
+ return __builtin_aarch64_fmav2sf (__b,
+ __aarch64_vdup_lane_f32 (__c, __lane),
+ __a);
+}
+
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfma_lane_f64 (float64_t __a, float64_t __b,
+ float64_t __c, const int __lane)
+{
+ return __builtin_fma (__b, __c, __a);
+}
+
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfmad_lane_f64 (float64_t __a, float64_t __b,
+ float64_t __c, const int __lane)
+{
+ return __builtin_fma (__b, __c, __a);
+}
+
+__extension__ static __inline float32_t __attribute__ ((__always_inline__))
+vfmas_lane_f32 (float32_t __a, float32_t __b,
+ float32x2_t __c, const int __lane)
+{
+ return __builtin_fmaf (__b, __aarch64_vget_lane_f32 (__c, __lane), __a);
+}
+
+/* vfma_laneq */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vfma_laneq_f32 (float32x2_t __a, float32x2_t __b,
+ float32x4_t __c, const int __lane)
+{
+ return __builtin_aarch64_fmav2sf (__b,
+ __aarch64_vdup_laneq_f32 (__c, __lane),
+ __a);
+}
+
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfma_laneq_f64 (float64_t __a, float64_t __b,
+ float64x2_t __c, const int __lane)
{
- return __aarch64_vget_lane_s16 (__a, __b);
+ return __builtin_fma (__b, __aarch64_vgetq_lane_f64 (__c, __lane), __a);
}
-__extension__ static __inline uint16_t __attribute__ ((__always_inline__))
-vduph_lane_u16 (uint16x4_t __a, const int __b)
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfmad_laneq_f64 (float64_t __a, float64_t __b,
+ float64x2_t __c, const int __lane)
{
- return __aarch64_vget_lane_u16 (__a, __b);
+ return __builtin_fma (__b, __aarch64_vgetq_lane_f64 (__c, __lane), __a);
}
-/* vdups_lane */
__extension__ static __inline float32_t __attribute__ ((__always_inline__))
-vdups_lane_f32 (float32x2_t __a, const int __b)
+vfmas_laneq_f32 (float32_t __a, float32_t __b,
+ float32x4_t __c, const int __lane)
{
- return __aarch64_vget_lane_f32 (__a, __b);
+ return __builtin_fmaf (__b, __aarch64_vgetq_lane_f32 (__c, __lane), __a);
}
-__extension__ static __inline int32_t __attribute__ ((__always_inline__))
-vdups_lane_s32 (int32x2_t __a, const int __b)
+/* vfmaq_lane */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vfmaq_lane_f32 (float32x4_t __a, float32x4_t __b,
+ float32x2_t __c, const int __lane)
{
- return __aarch64_vget_lane_s32 (__a, __b);
+ return __builtin_aarch64_fmav4sf (__b,
+ __aarch64_vdupq_lane_f32 (__c, __lane),
+ __a);
}
-__extension__ static __inline uint32_t __attribute__ ((__always_inline__))
-vdups_lane_u32 (uint32x2_t __a, const int __b)
+__extension__ static __inline float64x2_t __attribute__ ((__always_inline__))
+vfmaq_lane_f64 (float64x2_t __a, float64x2_t __b,
+ float64_t __c, const int __lane)
{
- return __aarch64_vget_lane_u32 (__a, __b);
+ return __builtin_aarch64_fmav2df (__b, vdupq_n_f64 (__c), __a);
}
-/* vdupd_lane */
-__extension__ static __inline float64_t __attribute__ ((__always_inline__))
-vdupd_lane_f64 (float64x1_t __a, const int __attribute__ ((unused)) __b)
+/* vfmaq_laneq */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vfmaq_laneq_f32 (float32x4_t __a, float32x4_t __b,
+ float32x4_t __c, const int __lane)
{
- return __a;
+ return __builtin_aarch64_fmav4sf (__b,
+ __aarch64_vdupq_laneq_f32 (__c, __lane),
+ __a);
}
-__extension__ static __inline int64_t __attribute__ ((__always_inline__))
-vdupd_lane_s64 (int64x1_t __a, const int __attribute__ ((unused)) __b)
+__extension__ static __inline float64x2_t __attribute__ ((__always_inline__))
+vfmaq_laneq_f64 (float64x2_t __a, float64x2_t __b,
+ float64x2_t __c, const int __lane)
{
- return __a;
+ return __builtin_aarch64_fmav2df (__b,
+ __aarch64_vdupq_laneq_f64 (__c, __lane),
+ __a);
}
-__extension__ static __inline uint64_t __attribute__ ((__always_inline__))
-vdupd_lane_u64 (uint64x1_t __a, const int __attribute__ ((unused)) __b)
+/* vfms_lane */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vfms_lane_f32 (float32x2_t __a, float32x2_t __b,
+ float32x2_t __c, const int __lane)
{
- return __a;
+ return __builtin_aarch64_fmav2sf (-__b,
+ __aarch64_vdup_lane_f32 (__c, __lane),
+ __a);
}
-/* vdupb_laneq */
-__extension__ static __inline poly8_t __attribute__ ((__always_inline__))
-vdupb_laneq_p8 (poly8x16_t __a, const int __b)
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfms_lane_f64 (float64_t __a, float64_t __b,
+ float64_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_p8 (__a, __b);
+ return __builtin_fma (-__b, __c, __a);
}
-__extension__ static __inline int8_t __attribute__ ((__always_inline__))
-vdupb_laneq_s8 (int8x16_t __a, const int __attribute__ ((unused)) __b)
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfmsd_lane_f64 (float64_t __a, float64_t __b,
+ float64_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_s8 (__a, __b);
+ return __builtin_fma (-__b, __c, __a);
}
-__extension__ static __inline uint8_t __attribute__ ((__always_inline__))
-vdupb_laneq_u8 (uint8x16_t __a, const int __b)
+__extension__ static __inline float32_t __attribute__ ((__always_inline__))
+vfmss_lane_f32 (float32_t __a, float32_t __b,
+ float32x2_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_u8 (__a, __b);
+ return __builtin_fmaf (-__b, __aarch64_vget_lane_f32 (__c, __lane), __a);
}
-/* vduph_laneq */
-__extension__ static __inline poly16_t __attribute__ ((__always_inline__))
-vduph_laneq_p16 (poly16x8_t __a, const int __b)
+/* vfms_laneq */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vfms_laneq_f32 (float32x2_t __a, float32x2_t __b,
+ float32x4_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_p16 (__a, __b);
+ return __builtin_aarch64_fmav2sf (-__b,
+ __aarch64_vdup_laneq_f32 (__c, __lane),
+ __a);
}
-__extension__ static __inline int16_t __attribute__ ((__always_inline__))
-vduph_laneq_s16 (int16x8_t __a, const int __b)
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfms_laneq_f64 (float64_t __a, float64_t __b,
+ float64x2_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_s16 (__a, __b);
+ return __builtin_fma (-__b, __aarch64_vgetq_lane_f64 (__c, __lane), __a);
}
-__extension__ static __inline uint16_t __attribute__ ((__always_inline__))
-vduph_laneq_u16 (uint16x8_t __a, const int __b)
+__extension__ static __inline float64_t __attribute__ ((__always_inline__))
+vfmsd_laneq_f64 (float64_t __a, float64_t __b,
+ float64x2_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_u16 (__a, __b);
+ return __builtin_fma (-__b, __aarch64_vgetq_lane_f64 (__c, __lane), __a);
}
-/* vdups_laneq */
__extension__ static __inline float32_t __attribute__ ((__always_inline__))
-vdups_laneq_f32 (float32x4_t __a, const int __b)
+vfmss_laneq_f32 (float32_t __a, float32_t __b,
+ float32x4_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_f32 (__a, __b);
+ return __builtin_fmaf (-__b, __aarch64_vgetq_lane_f32 (__c, __lane), __a);
}
-__extension__ static __inline int32_t __attribute__ ((__always_inline__))
-vdups_laneq_s32 (int32x4_t __a, const int __b)
-{
- return __aarch64_vgetq_lane_s32 (__a, __b);
-}
+/* vfmsq_lane */
-__extension__ static __inline uint32_t __attribute__ ((__always_inline__))
-vdups_laneq_u32 (uint32x4_t __a, const int __b)
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vfmsq_lane_f32 (float32x4_t __a, float32x4_t __b,
+ float32x2_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_u32 (__a, __b);
+ return __builtin_aarch64_fmav4sf (-__b,
+ __aarch64_vdupq_lane_f32 (__c, __lane),
+ __a);
}
-/* vdupd_laneq */
-__extension__ static __inline float64_t __attribute__ ((__always_inline__))
-vdupd_laneq_f64 (float64x2_t __a, const int __b)
+__extension__ static __inline float64x2_t __attribute__ ((__always_inline__))
+vfmsq_lane_f64 (float64x2_t __a, float64x2_t __b,
+ float64_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_f64 (__a, __b);
+ return __builtin_aarch64_fmav2df (-__b, vdupq_n_f64 (__c), __a);
}
-__extension__ static __inline int64_t __attribute__ ((__always_inline__))
-vdupd_laneq_s64 (int64x2_t __a, const int __b)
+/* vfmsq_laneq */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vfmsq_laneq_f32 (float32x4_t __a, float32x4_t __b,
+ float32x4_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_s64 (__a, __b);
+ return __builtin_aarch64_fmav4sf (-__b,
+ __aarch64_vdupq_laneq_f32 (__c, __lane),
+ __a);
}
-__extension__ static __inline uint64_t __attribute__ ((__always_inline__))
-vdupd_laneq_u64 (uint64x2_t __a, const int __b)
+__extension__ static __inline float64x2_t __attribute__ ((__always_inline__))
+vfmsq_laneq_f64 (float64x2_t __a, float64x2_t __b,
+ float64x2_t __c, const int __lane)
{
- return __aarch64_vgetq_lane_u64 (__a, __b);
+ return __builtin_aarch64_fmav2df (-__b,
+ __aarch64_vdupq_laneq_f64 (__c, __lane),
+ __a);
}
/* vld1 */
return a + b * c;
}
+/* vmla_lane */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vmla_lane_f32 (float32x2_t __a, float32x2_t __b,
+ float32x2_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x4_t __attribute__ ((__always_inline__))
+vmla_lane_s16 (int16x4_t __a, int16x4_t __b,
+ int16x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x2_t __attribute__ ((__always_inline__))
+vmla_lane_s32 (int32x2_t __a, int32x2_t __b,
+ int32x2_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x4_t __attribute__ ((__always_inline__))
+vmla_lane_u16 (uint16x4_t __a, uint16x4_t __b,
+ uint16x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x2_t __attribute__ ((__always_inline__))
+vmla_lane_u32 (uint32x2_t __a, uint32x2_t __b,
+ uint32x2_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_u32 (__c, __lane)));
+}
+
+/* vmla_laneq */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vmla_laneq_f32 (float32x2_t __a, float32x2_t __b,
+ float32x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x4_t __attribute__ ((__always_inline__))
+vmla_laneq_s16 (int16x4_t __a, int16x4_t __b,
+ int16x8_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x2_t __attribute__ ((__always_inline__))
+vmla_laneq_s32 (int32x2_t __a, int32x2_t __b,
+ int32x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x4_t __attribute__ ((__always_inline__))
+vmla_laneq_u16 (uint16x4_t __a, uint16x4_t __b,
+ uint16x8_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x2_t __attribute__ ((__always_inline__))
+vmla_laneq_u32 (uint32x2_t __a, uint32x2_t __b,
+ uint32x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_u32 (__c, __lane)));
+}
+
+/* vmlaq_lane */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vmlaq_lane_f32 (float32x4_t __a, float32x4_t __b,
+ float32x2_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x8_t __attribute__ ((__always_inline__))
+vmlaq_lane_s16 (int16x8_t __a, int16x8_t __b,
+ int16x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
+vmlaq_lane_s32 (int32x4_t __a, int32x4_t __b,
+ int32x2_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x8_t __attribute__ ((__always_inline__))
+vmlaq_lane_u16 (uint16x8_t __a, uint16x8_t __b,
+ uint16x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x4_t __attribute__ ((__always_inline__))
+vmlaq_lane_u32 (uint32x4_t __a, uint32x4_t __b,
+ uint32x2_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vget_lane_u32 (__c, __lane)));
+}
+
+ /* vmlaq_laneq */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vmlaq_laneq_f32 (float32x4_t __a, float32x4_t __b,
+ float32x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x8_t __attribute__ ((__always_inline__))
+vmlaq_laneq_s16 (int16x8_t __a, int16x8_t __b,
+ int16x8_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
+vmlaq_laneq_s32 (int32x4_t __a, int32x4_t __b,
+ int32x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x8_t __attribute__ ((__always_inline__))
+vmlaq_laneq_u16 (uint16x8_t __a, uint16x8_t __b,
+ uint16x8_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x4_t __attribute__ ((__always_inline__))
+vmlaq_laneq_u32 (uint32x4_t __a, uint32x4_t __b,
+ uint32x4_t __c, const int __lane)
+{
+ return (__a + (__b * __aarch64_vgetq_lane_u32 (__c, __lane)));
+}
+
+/* vmls */
+
__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
vmls_f32 (float32x2_t a, float32x2_t b, float32x2_t c)
{
return a - b * c;
}
+/* vmls_lane */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vmls_lane_f32 (float32x2_t __a, float32x2_t __b,
+ float32x2_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x4_t __attribute__ ((__always_inline__))
+vmls_lane_s16 (int16x4_t __a, int16x4_t __b,
+ int16x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x2_t __attribute__ ((__always_inline__))
+vmls_lane_s32 (int32x2_t __a, int32x2_t __b,
+ int32x2_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x4_t __attribute__ ((__always_inline__))
+vmls_lane_u16 (uint16x4_t __a, uint16x4_t __b,
+ uint16x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x2_t __attribute__ ((__always_inline__))
+vmls_lane_u32 (uint32x2_t __a, uint32x2_t __b,
+ uint32x2_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_u32 (__c, __lane)));
+}
+
+/* vmls_laneq */
+
+__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
+vmls_laneq_f32 (float32x2_t __a, float32x2_t __b,
+ float32x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x4_t __attribute__ ((__always_inline__))
+vmls_laneq_s16 (int16x4_t __a, int16x4_t __b,
+ int16x8_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x2_t __attribute__ ((__always_inline__))
+vmls_laneq_s32 (int32x2_t __a, int32x2_t __b,
+ int32x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x4_t __attribute__ ((__always_inline__))
+vmls_laneq_u16 (uint16x4_t __a, uint16x4_t __b,
+ uint16x8_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x2_t __attribute__ ((__always_inline__))
+vmls_laneq_u32 (uint32x2_t __a, uint32x2_t __b,
+ uint32x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_u32 (__c, __lane)));
+}
+
+/* vmlsq_lane */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vmlsq_lane_f32 (float32x4_t __a, float32x4_t __b,
+ float32x2_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x8_t __attribute__ ((__always_inline__))
+vmlsq_lane_s16 (int16x8_t __a, int16x8_t __b,
+ int16x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
+vmlsq_lane_s32 (int32x4_t __a, int32x4_t __b,
+ int32x2_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_s32 (__c, __lane)));
+}
+
+__extension__ static __inline uint16x8_t __attribute__ ((__always_inline__))
+vmlsq_lane_u16 (uint16x8_t __a, uint16x8_t __b,
+ uint16x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x4_t __attribute__ ((__always_inline__))
+vmlsq_lane_u32 (uint32x4_t __a, uint32x4_t __b,
+ uint32x2_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vget_lane_u32 (__c, __lane)));
+}
+
+ /* vmlsq_laneq */
+
+__extension__ static __inline float32x4_t __attribute__ ((__always_inline__))
+vmlsq_laneq_f32 (float32x4_t __a, float32x4_t __b,
+ float32x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_f32 (__c, __lane)));
+}
+
+__extension__ static __inline int16x8_t __attribute__ ((__always_inline__))
+vmlsq_laneq_s16 (int16x8_t __a, int16x8_t __b,
+ int16x8_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_s16 (__c, __lane)));
+}
+
+__extension__ static __inline int32x4_t __attribute__ ((__always_inline__))
+vmlsq_laneq_s32 (int32x4_t __a, int32x4_t __b,
+ int32x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_s32 (__c, __lane)));
+}
+__extension__ static __inline uint16x8_t __attribute__ ((__always_inline__))
+vmlsq_laneq_u16 (uint16x8_t __a, uint16x8_t __b,
+ uint16x8_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_u16 (__c, __lane)));
+}
+
+__extension__ static __inline uint32x4_t __attribute__ ((__always_inline__))
+vmlsq_laneq_u32 (uint32x4_t __a, uint32x4_t __b,
+ uint32x4_t __c, const int __lane)
+{
+ return (__a - (__b * __aarch64_vgetq_lane_u32 (__c, __lane)));
+}
+
/* vmul_lane */
__extension__ static __inline float32x2_t __attribute__ ((__always_inline__))
projects / gcc.git / commitdiff