swscale: Provide the right alignment for external mmx asm

[ffmpeg] / libavcodec / dsputil.h

diff --git a/libavcodec/dsputil.h b/libavcodec/dsputil.h
index 77980e02f8a8a2b3a04380ff205bb7513ccb02e9..bb70428e2aac8ed7899016b1832665052a44f150 100644 (file)
--- a/libavcodec/dsputil.h
+++ b/libavcodec/dsputil.h
@@ -101,18 +101,6 @@ PUTAVG_PIXELS(10)
 #define ff_put_pixels16x16_c ff_put_pixels16x16_8_c
 #define ff_avg_pixels16x16_c ff_avg_pixels16x16_8_c
 
-/* VP3 DSP functions */
-void ff_vp3_idct_c(DCTELEM *block/* align 16*/);
-void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
-void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
-void ff_vp3_idct_dc_add_c(uint8_t *dest/*align 8*/, int line_size, const DCTELEM *block/*align 16*/);
-
-void ff_vp3_v_loop_filter_c(uint8_t *src, int stride, int *bounding_values);
-void ff_vp3_h_loop_filter_c(uint8_t *src, int stride, int *bounding_values);
-
-/* EA functions */
-void ff_ea_idct_put_c(uint8_t *dest, int linesize, DCTELEM *block);
-
 /* RV40 functions */
 void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride);
 void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride);
@@ -209,10 +197,6 @@ EMULATED_EDGE(8)
 EMULATED_EDGE(9)
 EMULATED_EDGE(10)
 
-void ff_add_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int linesize);
-void ff_put_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int linesize);
-void ff_put_signed_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int linesize);
-
 /**
  * DSPContext.
  */
@@ -388,13 +372,6 @@ typedef struct DSPContext {
 
     void (*h261_loop_filter)(uint8_t *src, int stride);
 
-    void (*x8_v_loop_filter)(uint8_t *src, int stride, int qscale);
-    void (*x8_h_loop_filter)(uint8_t *src, int stride, int qscale);
-
-    void (*vp3_idct_dc_add)(uint8_t *dest/*align 8*/, int line_size, const DCTELEM *block/*align 16*/);
-    void (*vp3_v_loop_filter)(uint8_t *src, int stride, int *bounding_values);
-    void (*vp3_h_loop_filter)(uint8_t *src, int stride, int *bounding_values);
-
     /* assume len is a multiple of 4, and arrays are 16-byte aligned */
     void (*vorbis_inverse_coupling)(float *mag, float *ang, int blocksize);
     void (*ac3_downmix)(float (*samples)[256], float (*matrix)[2], int out_ch, int in_ch, int len);
@@ -509,11 +486,6 @@ typedef struct DSPContext {
                                unsigned int filter_shift, int32_t mask, int blocksize,
                                int32_t *sample_buffer);
 
-    /* intrax8 functions */
-    void (*x8_spatial_compensation[12])(uint8_t *src , uint8_t *dst, int linesize);
-    void (*x8_setup_spatial_compensation)(uint8_t *src, uint8_t *dst, int linesize,
-           int * range, int * sum,  int edges);
-
     /**
      * Calculate scalar product of two vectors.
      * @param len length of vectors, should be multiple of 16
@@ -565,6 +537,17 @@ void ff_dsputil_init(DSPContext* p, AVCodecContext *avctx);
 
 int ff_check_alignment(void);
 
+/**
+ * Return the scalar product of two vectors.
+ *
+ * @param v1  first input vector
+ * @param v2  first input vector
+ * @param len number of elements
+ *
+ * @return sum of elementwise products
+ */
+float ff_scalarproduct_float_c(const float *v1, const float *v2, int len);
+
 /**
  * permute block according to permuatation.
  * @param last last non zero element in scantable order
@@ -630,7 +613,6 @@ void ff_dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
 void ff_dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
 
 void ff_dsputil_init_dwt(DSPContext *c);
-void ff_intrax8dsp_init(DSPContext* c, AVCodecContext *avctx);
 void ff_mlp_init(DSPContext* c, AVCodecContext *avctx);
 void ff_mlp_init_x86(DSPContext* c, AVCodecContext *avctx);
 
@@ -640,22 +622,26 @@ void ff_mlp_init_x86(DSPContext* c, AVCodecContext *avctx);
 #   define STRIDE_ALIGN 8
 #endif
 
+// Some broken preprocessors need a second expansion
+// to be forced to tokenize __VA_ARGS__
+#define E(x) x
+
 #define LOCAL_ALIGNED_A(a, t, v, s, o, ...)             \
     uint8_t la_##v[sizeof(t s o) + (a)];                \
     t (*v) o = (void *)FFALIGN((uintptr_t)la_##v, a)
 
 #define LOCAL_ALIGNED_D(a, t, v, s, o, ...) DECLARE_ALIGNED(a, t, v) s o
 
-#define LOCAL_ALIGNED(a, t, v, ...) LOCAL_ALIGNED_A(a, t, v, __VA_ARGS__,,)
+#define LOCAL_ALIGNED(a, t, v, ...) E(LOCAL_ALIGNED_A(a, t, v, __VA_ARGS__,,))
 
 #if HAVE_LOCAL_ALIGNED_8
-#   define LOCAL_ALIGNED_8(t, v, ...) LOCAL_ALIGNED_D(8, t, v, __VA_ARGS__,,)
+#   define LOCAL_ALIGNED_8(t, v, ...) E(LOCAL_ALIGNED_D(8, t, v, __VA_ARGS__,,))
 #else
 #   define LOCAL_ALIGNED_8(t, v, ...) LOCAL_ALIGNED(8, t, v, __VA_ARGS__)
 #endif
 
 #if HAVE_LOCAL_ALIGNED_16
-#   define LOCAL_ALIGNED_16(t, v, ...) LOCAL_ALIGNED_D(16, t, v, __VA_ARGS__,,)
+#   define LOCAL_ALIGNED_16(t, v, ...) E(LOCAL_ALIGNED_D(16, t, v, __VA_ARGS__,,))
 #else
 #   define LOCAL_ALIGNED_16(t, v, ...) LOCAL_ALIGNED(16, t, v, __VA_ARGS__)
 #endif