Merge commit '73b704ac609d83e0be124589f24efd9b94947cf9'

* commit '73b704ac609d83e0be124589f24efd9b94947cf9':
  arm: Add some missing header #includes
  floatdsp: move scalarproduct_float from dsputil to avfloatdsp.

Conflicts:
	libavcodec/acelp_pitch_delay.c
	libavcodec/amrnbdec.c
	libavcodec/amrwbdec.c
	libavcodec/ra288.c
	libavcodec/x86/dsputil_mmx.c
	libavutil/x86/float_dsp.asm

Merged-by: Michael Niedermayer <michaelni@gmx.at>

6 files changed, 80 insertions, 1 deletions

diff --git a/libavutil/arm/float_dsp_init_neon.c b/libavutil/arm/float_dsp_init_neon.c
index b3644e82a2..a7245ad92b 100644
--- a/libavutil/arm/float_dsp_init_neon.c
+++ b/libavutil/arm/float_dsp_init_neon.c
@@ -43,6 +43,8 @@ void ff_vector_fmul_reverse_neon(float *dst, const float *src0,
 
 void ff_butterflies_float_neon(float *v1, float *v2, int len);
 
+float ff_scalarproduct_float_neon(const float *v1, const float *v2, int len);
+
 void ff_float_dsp_init_neon(AVFloatDSPContext *fdsp)
 {
     fdsp->vector_fmul = ff_vector_fmul_neon;
@@ -52,4 +54,5 @@ void ff_float_dsp_init_neon(AVFloatDSPContext *fdsp)
     fdsp->vector_fmul_add    = ff_vector_fmul_add_neon;
     fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_neon;
     fdsp->butterflies_float = ff_butterflies_float_neon;
+    fdsp->scalarproduct_float = ff_scalarproduct_float_neon;
 }
diff --git a/libavutil/arm/float_dsp_neon.S b/libavutil/arm/float_dsp_neon.S
index 4acc406d33..559b565628 100644
--- a/libavutil/arm/float_dsp_neon.S
+++ b/libavutil/arm/float_dsp_neon.S
@@ -256,3 +256,16 @@ function ff_butterflies_float_neon, export=1
         bgt             1b
         bx              lr
 endfunc
+
+function ff_scalarproduct_float_neon, export=1
+        vmov.f32        q2,  #0.0
+1:      vld1.32         {q0},[r0,:128]!
+        vld1.32         {q1},[r1,:128]!
+        vmla.f32        q2,  q0,  q1
+        subs            r2,  r2,  #4
+        bgt             1b
+        vadd.f32        d0,  d4,  d5
+        vpadd.f32       d0,  d0,  d0
+NOVFP   vmov.32         r0,  d0[0]
+        bx              lr
+endfunc
diff --git a/libavutil/float_dsp.c b/libavutil/float_dsp.c
index 05688e4c3f..50c9e60c32 100644
--- a/libavutil/float_dsp.c
+++ b/libavutil/float_dsp.c
@@ -104,6 +104,17 @@ static void butterflies_float_c(float *restrict v1, float *restrict v2,
     }
 }
 
+float avpriv_scalarproduct_float_c(const float *v1, const float *v2, int len)
+{
+    float p = 0.0;
+    int i;
+
+    for (i = 0; i < len; i++)
+        p += v1[i] * v2[i];
+
+    return p;
+}
+
 void avpriv_float_dsp_init(AVFloatDSPContext *fdsp, int bit_exact)
 {
     fdsp->vector_fmul = vector_fmul_c;
@@ -114,6 +125,7 @@ void avpriv_float_dsp_init(AVFloatDSPContext *fdsp, int bit_exact)
     fdsp->vector_fmul_add = vector_fmul_add_c;
     fdsp->vector_fmul_reverse = vector_fmul_reverse_c;
     fdsp->butterflies_float = butterflies_float_c;
+    fdsp->scalarproduct_float = avpriv_scalarproduct_float_c;
 
 #if ARCH_ARM
     ff_float_dsp_init_arm(fdsp);
diff --git a/libavutil/float_dsp.h b/libavutil/float_dsp.h
index ff83beddbe..6cc7e76c11 100644
--- a/libavutil/float_dsp.h
+++ b/libavutil/float_dsp.h
@@ -146,9 +146,31 @@ typedef struct AVFloatDSPContext {
      * @param len length of vectors, multiple of 4
      */
     void (*butterflies_float)(float *restrict v1, float *restrict v2, int len);
+
+    /**
+     * Calculate the scalar product of two vectors of floats.
+     *
+     * @param v1  first vector, 16-byte aligned
+     * @param v2  second vector, 16-byte aligned
+     * @param len length of vectors, multiple of 4
+     *
+     * @return sum of elementwise products
+     */
+    float (*scalarproduct_float)(const float *v1, const float *v2, int len);
 } AVFloatDSPContext;
 
 /**
+ * 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 avpriv_scalarproduct_float_c(const float *v1, const float *v2, int len);
+
+/**
  * Initialize a float DSP context.
  *
  * @param fdsp    float DSP context
diff --git a/libavutil/x86/float_dsp.asm b/libavutil/x86/float_dsp.asm
index 3e5e91ad07..004e6cf1fe 100644
--- a/libavutil/x86/float_dsp.asm
+++ b/libavutil/x86/float_dsp.asm
@@ -236,4 +236,30 @@ VECTOR_FMUL_REVERSE
 %if HAVE_AVX_EXTERNAL
 INIT_YMM avx
 VECTOR_FMUL_REVERSE
-%endif
\ No newline at end of file
+%endif
+
+; float scalarproduct_float_sse(const float *v1, const float *v2, int len)
+INIT_XMM sse
+cglobal scalarproduct_float, 3,3,2, v1, v2, offset
+    neg   offsetq
+    shl   offsetq, 2
+    sub       v1q, offsetq
+    sub       v2q, offsetq
+    xorps    xmm0, xmm0
+.loop:
+    movaps   xmm1, [v1q+offsetq]
+    mulps    xmm1, [v2q+offsetq]
+    addps    xmm0, xmm1
+    add   offsetq, 16
+    js .loop
+    movhlps  xmm1, xmm0
+    addps    xmm0, xmm1
+    movss    xmm1, xmm0
+    shufps   xmm0, xmm0, 1
+    addss    xmm0, xmm1
+%if ARCH_X86_64 == 0
+    movss     r0m,  xmm0
+    fld dword r0m
+%endif
+    RET
+
diff --git a/libavutil/x86/float_dsp_init.c b/libavutil/x86/float_dsp_init.c
index 9c58e2bc30..5c6383bc74 100644
--- a/libavutil/x86/float_dsp_init.c
+++ b/libavutil/x86/float_dsp_init.c
@@ -51,6 +51,8 @@ void ff_vector_fmul_reverse_sse(float *dst, const float *src0,
 void ff_vector_fmul_reverse_avx(float *dst, const float *src0,
                                 const float *src1, int len);
 
+float ff_scalarproduct_float_sse(const float *v1, const float *v2, int order);
+
 #if HAVE_6REGS && HAVE_INLINE_ASM
 static void vector_fmul_window_3dnowext(float *dst, const float *src0,
                                         const float *src1, const float *win,
@@ -135,6 +137,7 @@ void ff_float_dsp_init_x86(AVFloatDSPContext *fdsp)
         fdsp->vector_fmul_scalar = ff_vector_fmul_scalar_sse;
         fdsp->vector_fmul_add    = ff_vector_fmul_add_sse;
         fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_sse;
+        fdsp->scalarproduct_float = ff_scalarproduct_float_sse;
     }
     if (EXTERNAL_SSE2(mm_flags)) {
         fdsp->vector_dmul_scalar = ff_vector_dmul_scalar_sse2;