From 2fdbf70d4079342b9305bbbfcea2e632a565e739 Mon Sep 17 00:00:00 2001 From: Johann Date: Tue, 16 Dec 2014 15:08:16 -0800 Subject: [PATCH] Use defines for inline and __builtin_prefetch These were established for compatibility. Make sure to use them. Most frequently they manifest as issues on Visual Studio builds. Change-Id: I39d764d2eb341b999d7a6132cb44b2acfc511160 --- vp9/common/arm/neon/vp9_convolve8_avg_neon.c | 5 ++++- vp9/common/arm/neon/vp9_convolve8_neon.c | 5 ++++- vp9/common/arm/neon/vp9_idct16x16_add_neon.c | 4 +++- vp9/common/arm/neon/vp9_idct32x32_1_add_neon.c | 10 ++++++---- vp9/common/arm/neon/vp9_idct32x32_add_neon.c | 14 ++++++++------ vp9/common/arm/neon/vp9_idct8x8_add_neon.c | 6 ++++-- vp9/common/arm/neon/vp9_iht4x4_add_neon.c | 11 ++++++----- vp9/common/arm/neon/vp9_iht8x8_add_neon.c | 7 ++++--- vp9/common/arm/neon/vp9_loopfilter_neon.c | 6 ++++-- 9 files changed, 43 insertions(+), 25 deletions(-) diff --git a/vp9/common/arm/neon/vp9_convolve8_avg_neon.c b/vp9/common/arm/neon/vp9_convolve8_avg_neon.c index 2f8dda0..dd569d3 100644 --- a/vp9/common/arm/neon/vp9_convolve8_avg_neon.c +++ b/vp9/common/arm/neon/vp9_convolve8_avg_neon.c @@ -11,6 +11,9 @@ #include #include +#include "./vpx_config.h" +#include "vpx_ports/mem.h" + void vp9_convolve8_avg_horiz_c(const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, @@ -22,7 +25,7 @@ void vp9_convolve8_avg_vert_c(const uint8_t *src, ptrdiff_t src_stride, const int16_t *filter_y, int y_step_q4, int w, int h); -static inline int32x4_t MULTIPLY_BY_Q0( +static INLINE int32x4_t MULTIPLY_BY_Q0( int16x4_t dsrc0, int16x4_t dsrc1, int16x4_t dsrc2, diff --git a/vp9/common/arm/neon/vp9_convolve8_neon.c b/vp9/common/arm/neon/vp9_convolve8_neon.c index c8704aa..5c555c4 100644 --- a/vp9/common/arm/neon/vp9_convolve8_neon.c +++ b/vp9/common/arm/neon/vp9_convolve8_neon.c @@ -11,6 +11,9 @@ #include #include +#include "./vpx_config.h" +#include "vpx_ports/mem.h" + void vp9_convolve8_horiz_c(const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, @@ -22,7 +25,7 @@ void vp9_convolve8_vert_c(const uint8_t *src, ptrdiff_t src_stride, const int16_t *filter_y, int y_step_q4, int w, int h); -static inline int32x4_t MULTIPLY_BY_Q0( +static INLINE int32x4_t MULTIPLY_BY_Q0( int16x4_t dsrc0, int16x4_t dsrc1, int16x4_t dsrc2, diff --git a/vp9/common/arm/neon/vp9_idct16x16_add_neon.c b/vp9/common/arm/neon/vp9_idct16x16_add_neon.c index 68d7ccc..5fa3f5c 100644 --- a/vp9/common/arm/neon/vp9_idct16x16_add_neon.c +++ b/vp9/common/arm/neon/vp9_idct16x16_add_neon.c @@ -10,6 +10,8 @@ #include +#include "./vpx_config.h" + static int16_t cospi_2_64 = 16305; static int16_t cospi_4_64 = 16069; static int16_t cospi_6_64 = 15679; @@ -26,7 +28,7 @@ static int16_t cospi_26_64 = 4756; static int16_t cospi_28_64 = 3196; static int16_t cospi_30_64 = 1606; -static inline void TRANSPOSE8X8( +static INLINE void TRANSPOSE8X8( int16x8_t *q8s16, int16x8_t *q9s16, int16x8_t *q10s16, diff --git a/vp9/common/arm/neon/vp9_idct32x32_1_add_neon.c b/vp9/common/arm/neon/vp9_idct32x32_1_add_neon.c index 1bfee22..d0e4b4f 100644 --- a/vp9/common/arm/neon/vp9_idct32x32_1_add_neon.c +++ b/vp9/common/arm/neon/vp9_idct32x32_1_add_neon.c @@ -11,7 +11,9 @@ #include #include "vp9/common/vp9_idct.h" -static inline void LD_16x8( +#include "./vpx_config.h" + +static INLINE void LD_16x8( uint8_t *d, int d_stride, uint8x16_t *q8u8, @@ -40,7 +42,7 @@ static inline void LD_16x8( return; } -static inline void ADD_DIFF_16x8( +static INLINE void ADD_DIFF_16x8( uint8x16_t qdiffu8, uint8x16_t *q8u8, uint8x16_t *q9u8, @@ -61,7 +63,7 @@ static inline void ADD_DIFF_16x8( return; } -static inline void SUB_DIFF_16x8( +static INLINE void SUB_DIFF_16x8( uint8x16_t qdiffu8, uint8x16_t *q8u8, uint8x16_t *q9u8, @@ -82,7 +84,7 @@ static inline void SUB_DIFF_16x8( return; } -static inline void ST_16x8( +static INLINE void ST_16x8( uint8_t *d, int d_stride, uint8x16_t *q8u8, diff --git a/vp9/common/arm/neon/vp9_idct32x32_add_neon.c b/vp9/common/arm/neon/vp9_idct32x32_add_neon.c index 53f721b..309bdf8 100644 --- a/vp9/common/arm/neon/vp9_idct32x32_add_neon.c +++ b/vp9/common/arm/neon/vp9_idct32x32_add_neon.c @@ -10,6 +10,8 @@ #include +#include "./vpx_config.h" + static int16_t cospi_1_64 = 16364; static int16_t cospi_2_64 = 16305; static int16_t cospi_3_64 = 16207; @@ -57,7 +59,7 @@ static int16_t cospi_31_64 = 804; #define STORE_COMBINE_CENTER_RESULTS(r10, r9) \ __STORE_COMBINE_CENTER_RESULTS(r10, r9, stride, \ q6s16, q7s16, q8s16, q9s16); -static inline void __STORE_COMBINE_CENTER_RESULTS( +static INLINE void __STORE_COMBINE_CENTER_RESULTS( uint8_t *p1, uint8_t *p2, int stride, @@ -105,7 +107,7 @@ static inline void __STORE_COMBINE_CENTER_RESULTS( #define STORE_COMBINE_EXTREME_RESULTS(r7, r6); \ __STORE_COMBINE_EXTREME_RESULTS(r7, r6, stride, \ q4s16, q5s16, q6s16, q7s16); -static inline void __STORE_COMBINE_EXTREME_RESULTS( +static INLINE void __STORE_COMBINE_EXTREME_RESULTS( uint8_t *p1, uint8_t *p2, int stride, @@ -152,7 +154,7 @@ static inline void __STORE_COMBINE_EXTREME_RESULTS( #define DO_BUTTERFLY_STD(const_1, const_2, qA, qB) \ DO_BUTTERFLY(q14s16, q13s16, const_1, const_2, qA, qB); -static inline void DO_BUTTERFLY( +static INLINE void DO_BUTTERFLY( int16x8_t q14s16, int16x8_t q13s16, int16_t first_const, @@ -194,7 +196,7 @@ static inline void DO_BUTTERFLY( return; } -static inline void idct32_transpose_pair( +static INLINE void idct32_transpose_pair( int16_t *input, int16_t *t_buf) { int16_t *in; @@ -288,7 +290,7 @@ static inline void idct32_transpose_pair( return; } -static inline void idct32_bands_end_1st_pass( +static INLINE void idct32_bands_end_1st_pass( int16_t *out, int16x8_t q2s16, int16x8_t q3s16, @@ -383,7 +385,7 @@ static inline void idct32_bands_end_1st_pass( return; } -static inline void idct32_bands_end_2nd_pass( +static INLINE void idct32_bands_end_2nd_pass( int16_t *out, uint8_t *dest, int stride, diff --git a/vp9/common/arm/neon/vp9_idct8x8_add_neon.c b/vp9/common/arm/neon/vp9_idct8x8_add_neon.c index 50587f6..2b3c1ce 100644 --- a/vp9/common/arm/neon/vp9_idct8x8_add_neon.c +++ b/vp9/common/arm/neon/vp9_idct8x8_add_neon.c @@ -10,6 +10,8 @@ #include +#include "./vpx_config.h" + static int16_t cospi_4_64 = 16069; static int16_t cospi_8_64 = 15137; static int16_t cospi_12_64 = 13623; @@ -18,7 +20,7 @@ static int16_t cospi_20_64 = 9102; static int16_t cospi_24_64 = 6270; static int16_t cospi_28_64 = 3196; -static inline void TRANSPOSE8X8( +static INLINE void TRANSPOSE8X8( int16x8_t *q8s16, int16x8_t *q9s16, int16x8_t *q10s16, @@ -87,7 +89,7 @@ static inline void TRANSPOSE8X8( return; } -static inline void IDCT8x8_1D( +static INLINE void IDCT8x8_1D( int16x8_t *q8s16, int16x8_t *q9s16, int16x8_t *q10s16, diff --git a/vp9/common/arm/neon/vp9_iht4x4_add_neon.c b/vp9/common/arm/neon/vp9_iht4x4_add_neon.c index cd8c358..1761fad 100644 --- a/vp9/common/arm/neon/vp9_iht4x4_add_neon.c +++ b/vp9/common/arm/neon/vp9_iht4x4_add_neon.c @@ -12,6 +12,7 @@ #include #include "./vp9_rtcd.h" +#include "./vpx_config.h" #include "vp9/common/vp9_common.h" static int16_t sinpi_1_9 = 0x14a3; @@ -22,7 +23,7 @@ static int16_t cospi_8_64 = 0x3b21; static int16_t cospi_16_64 = 0x2d41; static int16_t cospi_24_64 = 0x187e; -static inline void TRANSPOSE4X4( +static INLINE void TRANSPOSE4X4( int16x8_t *q8s16, int16x8_t *q9s16) { int32x4_t q8s32, q9s32; @@ -41,7 +42,7 @@ static inline void TRANSPOSE4X4( return; } -static inline void GENERATE_COSINE_CONSTANTS( +static INLINE void GENERATE_COSINE_CONSTANTS( int16x4_t *d0s16, int16x4_t *d1s16, int16x4_t *d2s16) { @@ -51,7 +52,7 @@ static inline void GENERATE_COSINE_CONSTANTS( return; } -static inline void GENERATE_SINE_CONSTANTS( +static INLINE void GENERATE_SINE_CONSTANTS( int16x4_t *d3s16, int16x4_t *d4s16, int16x4_t *d5s16, @@ -63,7 +64,7 @@ static inline void GENERATE_SINE_CONSTANTS( return; } -static inline void IDCT4x4_1D( +static INLINE void IDCT4x4_1D( int16x4_t *d0s16, int16x4_t *d1s16, int16x4_t *d2s16, @@ -103,7 +104,7 @@ static inline void IDCT4x4_1D( return; } -static inline void IADST4x4_1D( +static INLINE void IADST4x4_1D( int16x4_t *d3s16, int16x4_t *d4s16, int16x4_t *d5s16, diff --git a/vp9/common/arm/neon/vp9_iht8x8_add_neon.c b/vp9/common/arm/neon/vp9_iht8x8_add_neon.c index 03c836d..04b342c 100644 --- a/vp9/common/arm/neon/vp9_iht8x8_add_neon.c +++ b/vp9/common/arm/neon/vp9_iht8x8_add_neon.c @@ -12,6 +12,7 @@ #include #include "./vp9_rtcd.h" +#include "./vpx_config.h" #include "vp9/common/vp9_common.h" static int16_t cospi_2_64 = 16305; @@ -30,7 +31,7 @@ static int16_t cospi_26_64 = 4756; static int16_t cospi_28_64 = 3196; static int16_t cospi_30_64 = 1606; -static inline void TRANSPOSE8X8( +static INLINE void TRANSPOSE8X8( int16x8_t *q8s16, int16x8_t *q9s16, int16x8_t *q10s16, @@ -99,7 +100,7 @@ static inline void TRANSPOSE8X8( return; } -static inline void IDCT8x8_1D( +static INLINE void IDCT8x8_1D( int16x8_t *q8s16, int16x8_t *q9s16, int16x8_t *q10s16, @@ -255,7 +256,7 @@ static inline void IDCT8x8_1D( return; } -static inline void IADST8X8_1D( +static INLINE void IADST8X8_1D( int16x8_t *q8s16, int16x8_t *q9s16, int16x8_t *q10s16, diff --git a/vp9/common/arm/neon/vp9_loopfilter_neon.c b/vp9/common/arm/neon/vp9_loopfilter_neon.c index f54d7a9..079d266 100644 --- a/vp9/common/arm/neon/vp9_loopfilter_neon.c +++ b/vp9/common/arm/neon/vp9_loopfilter_neon.c @@ -10,7 +10,9 @@ #include -static inline void vp9_loop_filter_neon( +#include "./vpx_config.h" + +static INLINE void vp9_loop_filter_neon( uint8x8_t dblimit, // flimit uint8x8_t dlimit, // limit uint8x8_t dthresh, // thresh @@ -271,7 +273,7 @@ void vp9_lpf_vertical_4_neon( return; } -static inline void vp9_mbloop_filter_neon( +static INLINE void vp9_mbloop_filter_neon( uint8x8_t dblimit, // mblimit uint8x8_t dlimit, // limit uint8x8_t dthresh, // thresh -- 2.7.4