Index: source/config/linux/arm-neon-cpu-detect/vp9_rtcd.h |
diff --git a/source/config/linux/arm-neon-cpu-detect/vp9_rtcd.h b/source/config/linux/arm-neon-cpu-detect/vp9_rtcd.h |
index 1ae87d1eba9bc09373a7b47fe4899ee1dfe000f0..1173e9bb72789a632675c634ec0c6e3cf5dbe1e8 100644 |
--- a/source/config/linux/arm-neon-cpu-detect/vp9_rtcd.h |
+++ b/source/config/linux/arm-neon-cpu-detect/vp9_rtcd.h |
@@ -33,7 +33,8 @@ unsigned int vp9_avg_4x4_c(const uint8_t *, int p); |
#define vp9_avg_4x4 vp9_avg_4x4_c |
unsigned int vp9_avg_8x8_c(const uint8_t *, int p); |
-#define vp9_avg_8x8 vp9_avg_8x8_c |
+unsigned int vp9_avg_8x8_neon(const uint8_t *, int p); |
+RTCD_EXTERN unsigned int (*vp9_avg_8x8)(const uint8_t *, int p); |
int64_t vp9_block_error_c(const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz); |
#define vp9_block_error vp9_block_error_c |
@@ -226,7 +227,8 @@ void vp9_fdct8x8_1_neon(const int16_t *input, tran_low_t *output, int stride); |
RTCD_EXTERN void (*vp9_fdct8x8_1)(const int16_t *input, tran_low_t *output, int stride); |
void vp9_fdct8x8_quant_c(const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan); |
-#define vp9_fdct8x8_quant vp9_fdct8x8_quant_c |
+void vp9_fdct8x8_quant_neon(const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan); |
+RTCD_EXTERN void (*vp9_fdct8x8_quant)(const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan); |
void vp9_fht16x16_c(const int16_t *input, tran_low_t *output, int stride, int tx_type); |
#define vp9_fht16x16 vp9_fht16x16_c |
@@ -414,7 +416,8 @@ void vp9_sad16x16x3_c(const uint8_t *src_ptr, int source_stride, const uint8_t * |
#define vp9_sad16x16x3 vp9_sad16x16x3_c |
void vp9_sad16x16x4d_c(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
-#define vp9_sad16x16x4d vp9_sad16x16x4d_c |
+void vp9_sad16x16x4d_neon(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
+RTCD_EXTERN void (*vp9_sad16x16x4d)(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
void vp9_sad16x16x8_c(const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array); |
#define vp9_sad16x16x8 vp9_sad16x16x8_c |
@@ -463,7 +466,8 @@ void vp9_sad32x32x3_c(const uint8_t *src_ptr, int source_stride, const uint8_t * |
#define vp9_sad32x32x3 vp9_sad32x32x3_c |
void vp9_sad32x32x4d_c(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
-#define vp9_sad32x32x4d vp9_sad32x32x4d_c |
+void vp9_sad32x32x4d_neon(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
+RTCD_EXTERN void (*vp9_sad32x32x4d)(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
void vp9_sad32x32x8_c(const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array); |
#define vp9_sad32x32x8 vp9_sad32x32x8_c |
@@ -524,7 +528,8 @@ void vp9_sad64x64x3_c(const uint8_t *src_ptr, int source_stride, const uint8_t * |
#define vp9_sad64x64x3 vp9_sad64x64x3_c |
void vp9_sad64x64x4d_c(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
-#define vp9_sad64x64x4d vp9_sad64x64x4d_c |
+void vp9_sad64x64x4d_neon(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
+RTCD_EXTERN void (*vp9_sad64x64x4d)(const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array); |
void vp9_sad64x64x8_c(const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array); |
#define vp9_sad64x64x8 vp9_sad64x64x8_c |
@@ -750,6 +755,8 @@ static void setup_rtcd_internal(void) |
(void)flags; |
+ vp9_avg_8x8 = vp9_avg_8x8_c; |
+ if (flags & HAS_NEON) vp9_avg_8x8 = vp9_avg_8x8_neon; |
vp9_convolve8 = vp9_convolve8_c; |
if (flags & HAS_NEON) vp9_convolve8 = vp9_convolve8_neon; |
vp9_convolve8_avg = vp9_convolve8_avg_c; |
@@ -770,6 +777,8 @@ static void setup_rtcd_internal(void) |
if (flags & HAS_NEON) vp9_fdct8x8 = vp9_fdct8x8_neon; |
vp9_fdct8x8_1 = vp9_fdct8x8_1_c; |
if (flags & HAS_NEON) vp9_fdct8x8_1 = vp9_fdct8x8_1_neon; |
+ vp9_fdct8x8_quant = vp9_fdct8x8_quant_c; |
+ if (flags & HAS_NEON) vp9_fdct8x8_quant = vp9_fdct8x8_quant_neon; |
vp9_get16x16var = vp9_get16x16var_c; |
if (flags & HAS_NEON) vp9_get16x16var = vp9_get16x16var_neon; |
vp9_get8x8var = vp9_get8x8var_c; |
@@ -834,10 +843,16 @@ static void setup_rtcd_internal(void) |
if (flags & HAS_NEON) vp9_quantize_fp = vp9_quantize_fp_neon; |
vp9_sad16x16 = vp9_sad16x16_c; |
if (flags & HAS_NEON) vp9_sad16x16 = vp9_sad16x16_neon; |
+ vp9_sad16x16x4d = vp9_sad16x16x4d_c; |
+ if (flags & HAS_NEON) vp9_sad16x16x4d = vp9_sad16x16x4d_neon; |
vp9_sad32x32 = vp9_sad32x32_c; |
if (flags & HAS_NEON) vp9_sad32x32 = vp9_sad32x32_neon; |
+ vp9_sad32x32x4d = vp9_sad32x32x4d_c; |
+ if (flags & HAS_NEON) vp9_sad32x32x4d = vp9_sad32x32x4d_neon; |
vp9_sad64x64 = vp9_sad64x64_c; |
if (flags & HAS_NEON) vp9_sad64x64 = vp9_sad64x64_neon; |
+ vp9_sad64x64x4d = vp9_sad64x64x4d_c; |
+ if (flags & HAS_NEON) vp9_sad64x64x4d = vp9_sad64x64x4d_neon; |
vp9_sad8x8 = vp9_sad8x8_c; |
if (flags & HAS_NEON) vp9_sad8x8 = vp9_sad8x8_neon; |
vp9_sub_pixel_variance16x16 = vp9_sub_pixel_variance16x16_c; |