Merge commit '9a9e2f1c8aa4539a261625145e5c1f46a8106ac2'

* commit '9a9e2f1c8aa4539a261625145e5c1f46a8106ac2':
  dsputil: Split audio operations off into a separate context

Conflicts:
	configure
	libavcodec/takdec.c
	libavcodec/x86/Makefile
	libavcodec/x86/dsputil.asm
	libavcodec/x86/dsputil_init.c
	libavcodec/x86/dsputil_mmx.c
	libavcodec/x86/dsputil_x86.h

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

8 files changed, 226 insertions, 153 deletions

diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
index fa03f7ce26..a08c525e30 100644
--- a/libavcodec/x86/Makefile
+++ b/libavcodec/x86/Makefile
@@ -2,6 +2,7 @@ OBJS                                   += x86/constants.o               \
                                           x86/fmtconvert_init.o         \
 
 OBJS-$(CONFIG_AC3DSP)                  += x86/ac3dsp_init.o
+OBJS-$(CONFIG_AUDIODSP)                += x86/audiodsp_init.o
 OBJS-$(CONFIG_BLOCKDSP)                += x86/blockdsp_mmx.o
 OBJS-$(CONFIG_DCT)                     += x86/dct_init.o
 OBJS-$(CONFIG_DSPUTIL)                 += x86/dsputil_init.o
@@ -69,6 +70,7 @@ YASM-OBJS                              += x86/deinterlace.o             \
                                           x86/fmtconvert.o              \
 
 YASM-OBJS-$(CONFIG_AC3DSP)             += x86/ac3dsp.o
+YASM-OBJS-$(CONFIG_AUDIODSP)           += x86/audiodsp.o
 YASM-OBJS-$(CONFIG_BLOCKDSP)           += x86/blockdsp.o
 YASM-OBJS-$(CONFIG_DCT)                += x86/dct32.o
 YASM-OBJS-$(CONFIG_DIRAC_DECODER)      += x86/diracdsp_mmx.o x86/diracdsp_yasm.o\
diff --git a/libavcodec/x86/audiodsp.asm b/libavcodec/x86/audiodsp.asm
new file mode 100644
index 0000000000..83f9bb6f45
--- /dev/null
+++ b/libavcodec/x86/audiodsp.asm
@@ -0,0 +1,133 @@
+;******************************************************************************
+;* optimized audio functions
+;* Copyright (c) 2008 Loren Merritt
+;*
+;* This file is part of FFmpeg.
+;*
+;* FFmpeg is free software; you can redistribute it and/or
+;* modify it under the terms of the GNU Lesser General Public
+;* License as published by the Free Software Foundation; either
+;* version 2.1 of the License, or (at your option) any later version.
+;*
+;* FFmpeg is distributed in the hope that it will be useful,
+;* but WITHOUT ANY WARRANTY; without even the implied warranty of
+;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+;* Lesser General Public License for more details.
+;*
+;* You should have received a copy of the GNU Lesser General Public
+;* License along with FFmpeg; if not, write to the Free Software
+;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+;******************************************************************************
+
+%include "libavutil/x86/x86util.asm"
+
+SECTION_TEXT
+
+%macro SCALARPRODUCT 0
+; int ff_scalarproduct_int16(int16_t *v1, int16_t *v2, int order)
+cglobal scalarproduct_int16, 3,3,3, v1, v2, order
+    shl orderq, 1
+    add v1q, orderq
+    add v2q, orderq
+    neg orderq
+    pxor    m2, m2
+.loop:
+    movu    m0, [v1q + orderq]
+    movu    m1, [v1q + orderq + mmsize]
+    pmaddwd m0, [v2q + orderq]
+    pmaddwd m1, [v2q + orderq + mmsize]
+    paddd   m2, m0
+    paddd   m2, m1
+    add     orderq, mmsize*2
+    jl .loop
+    HADDD   m2, m0
+    movd   eax, m2
+%if mmsize == 8
+    emms
+%endif
+    RET
+%endmacro
+
+INIT_MMX mmxext
+SCALARPRODUCT
+INIT_XMM sse2
+SCALARPRODUCT
+
+
+;-----------------------------------------------------------------------------
+; void ff_vector_clip_int32(int32_t *dst, const int32_t *src, int32_t min,
+;                           int32_t max, unsigned int len)
+;-----------------------------------------------------------------------------
+
+; %1 = number of xmm registers used
+; %2 = number of inline load/process/store loops per asm loop
+; %3 = process 4*mmsize (%3=0) or 8*mmsize (%3=1) bytes per loop
+; %4 = CLIPD function takes min/max as float instead of int (CLIPD_SSE2)
+; %5 = suffix
+%macro VECTOR_CLIP_INT32 4-5
+cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max, len
+%if %4
+    cvtsi2ss  m4, minm
+    cvtsi2ss  m5, maxm
+%else
+    movd      m4, minm
+    movd      m5, maxm
+%endif
+    SPLATD    m4
+    SPLATD    m5
+.loop:
+%assign %%i 0
+%rep %2
+    mova      m0,  [srcq+mmsize*(0+%%i)]
+    mova      m1,  [srcq+mmsize*(1+%%i)]
+    mova      m2,  [srcq+mmsize*(2+%%i)]
+    mova      m3,  [srcq+mmsize*(3+%%i)]
+%if %3
+    mova      m7,  [srcq+mmsize*(4+%%i)]
+    mova      m8,  [srcq+mmsize*(5+%%i)]
+    mova      m9,  [srcq+mmsize*(6+%%i)]
+    mova      m10, [srcq+mmsize*(7+%%i)]
+%endif
+    CLIPD  m0,  m4, m5, m6
+    CLIPD  m1,  m4, m5, m6
+    CLIPD  m2,  m4, m5, m6
+    CLIPD  m3,  m4, m5, m6
+%if %3
+    CLIPD  m7,  m4, m5, m6
+    CLIPD  m8,  m4, m5, m6
+    CLIPD  m9,  m4, m5, m6
+    CLIPD  m10, m4, m5, m6
+%endif
+    mova  [dstq+mmsize*(0+%%i)], m0
+    mova  [dstq+mmsize*(1+%%i)], m1
+    mova  [dstq+mmsize*(2+%%i)], m2
+    mova  [dstq+mmsize*(3+%%i)], m3
+%if %3
+    mova  [dstq+mmsize*(4+%%i)], m7
+    mova  [dstq+mmsize*(5+%%i)], m8
+    mova  [dstq+mmsize*(6+%%i)], m9
+    mova  [dstq+mmsize*(7+%%i)], m10
+%endif
+%assign %%i %%i+4*(%3+1)
+%endrep
+    add     srcq, mmsize*4*(%2+%3)
+    add     dstq, mmsize*4*(%2+%3)
+    sub     lend, mmsize*(%2+%3)
+    jg .loop
+    REP_RET
+%endmacro
+
+INIT_MMX mmx
+%define CLIPD CLIPD_MMX
+VECTOR_CLIP_INT32 0, 1, 0, 0
+INIT_XMM sse2
+VECTOR_CLIP_INT32 6, 1, 0, 0, _int
+%define CLIPD CLIPD_SSE2
+VECTOR_CLIP_INT32 6, 2, 0, 1
+INIT_XMM sse4
+%define CLIPD CLIPD_SSE41
+%ifdef m8
+VECTOR_CLIP_INT32 11, 1, 1, 0
+%else
+VECTOR_CLIP_INT32 6, 1, 0, 0
+%endif
diff --git a/libavcodec/x86/audiodsp.h b/libavcodec/x86/audiodsp.h
new file mode 100644
index 0000000000..35f9f1485b
--- /dev/null
+++ b/libavcodec/x86/audiodsp.h
@@ -0,0 +1,25 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVCODEC_X86_AUDIODSP_H
+#define AVCODEC_X86_AUDIODSP_H
+
+void ff_vector_clipf_sse(float *dst, const float *src,
+                         float min, float max, int len);
+
+#endif /* AVCODEC_X86_AUDIODSP_H */
diff --git a/libavcodec/x86/audiodsp_init.c b/libavcodec/x86/audiodsp_init.c
new file mode 100644
index 0000000000..d586bf6c04
--- /dev/null
+++ b/libavcodec/x86/audiodsp_init.c
@@ -0,0 +1,66 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <stdint.h>
+
+#include "config.h"
+#include "libavutil/attributes.h"
+#include "libavutil/cpu.h"
+#include "libavutil/x86/asm.h"
+#include "libavutil/x86/cpu.h"
+#include "libavcodec/audiodsp.h"
+#include "audiodsp.h"
+
+int32_t ff_scalarproduct_int16_mmxext(const int16_t *v1, const int16_t *v2,
+                                      int order);
+int32_t ff_scalarproduct_int16_sse2(const int16_t *v1, const int16_t *v2,
+                                    int order);
+
+void ff_vector_clip_int32_mmx(int32_t *dst, const int32_t *src,
+                              int32_t min, int32_t max, unsigned int len);
+void ff_vector_clip_int32_sse2(int32_t *dst, const int32_t *src,
+                               int32_t min, int32_t max, unsigned int len);
+void ff_vector_clip_int32_int_sse2(int32_t *dst, const int32_t *src,
+                                   int32_t min, int32_t max, unsigned int len);
+void ff_vector_clip_int32_sse4(int32_t *dst, const int32_t *src,
+                               int32_t min, int32_t max, unsigned int len);
+
+av_cold void ff_audiodsp_init_x86(AudioDSPContext *c)
+{
+    int cpu_flags = av_get_cpu_flags();
+
+    if (EXTERNAL_MMX(cpu_flags))
+        c->vector_clip_int32 = ff_vector_clip_int32_mmx;
+
+    if (EXTERNAL_MMXEXT(cpu_flags))
+        c->scalarproduct_int16 = ff_scalarproduct_int16_mmxext;
+
+    if (EXTERNAL_SSE(cpu_flags))
+        c->vector_clipf = ff_vector_clipf_sse;
+
+    if (EXTERNAL_SSE2(cpu_flags)) {
+        c->scalarproduct_int16 = ff_scalarproduct_int16_sse2;
+        if (cpu_flags & AV_CPU_FLAG_ATOM)
+            c->vector_clip_int32 = ff_vector_clip_int32_int_sse2;
+        else
+            c->vector_clip_int32 = ff_vector_clip_int32_sse2;
+    }
+
+    if (EXTERNAL_SSE4(cpu_flags))
+        c->vector_clip_int32 = ff_vector_clip_int32_sse4;
+}
diff --git a/libavcodec/x86/dsputil.asm b/libavcodec/x86/dsputil.asm
index 3bb5d9cbfe..e261c0fcc7 100644
--- a/libavcodec/x86/dsputil.asm
+++ b/libavcodec/x86/dsputil.asm
@@ -30,115 +30,6 @@ cextern pb_80
 
 SECTION_TEXT
 
-%macro SCALARPRODUCT 0
-; int ff_scalarproduct_int16(int16_t *v1, int16_t *v2, int order)
-cglobal scalarproduct_int16, 3,3,3, v1, v2, order
-    shl orderq, 1
-    add v1q, orderq
-    add v2q, orderq
-    neg orderq
-    pxor    m2, m2
-.loop:
-    movu    m0, [v1q + orderq]
-    movu    m1, [v1q + orderq + mmsize]
-    pmaddwd m0, [v2q + orderq]
-    pmaddwd m1, [v2q + orderq + mmsize]
-    paddd   m2, m0
-    paddd   m2, m1
-    add     orderq, mmsize*2
-    jl .loop
-    HADDD   m2, m0
-    movd   eax, m2
-%if mmsize == 8
-    emms
-%endif
-    RET
-%endmacro
-
-INIT_MMX mmxext
-SCALARPRODUCT
-INIT_XMM sse2
-SCALARPRODUCT
-
-
-;-----------------------------------------------------------------------------
-; void ff_vector_clip_int32(int32_t *dst, const int32_t *src, int32_t min,
-;                           int32_t max, unsigned int len)
-;-----------------------------------------------------------------------------
-
-; %1 = number of xmm registers used
-; %2 = number of inline load/process/store loops per asm loop
-; %3 = process 4*mmsize (%3=0) or 8*mmsize (%3=1) bytes per loop
-; %4 = CLIPD function takes min/max as float instead of int (CLIPD_SSE2)
-; %5 = suffix
-%macro VECTOR_CLIP_INT32 4-5
-cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max, len
-%if %4
-    cvtsi2ss  m4, minm
-    cvtsi2ss  m5, maxm
-%else
-    movd      m4, minm
-    movd      m5, maxm
-%endif
-    SPLATD    m4
-    SPLATD    m5
-.loop:
-%assign %%i 0
-%rep %2
-    mova      m0,  [srcq+mmsize*(0+%%i)]
-    mova      m1,  [srcq+mmsize*(1+%%i)]
-    mova      m2,  [srcq+mmsize*(2+%%i)]
-    mova      m3,  [srcq+mmsize*(3+%%i)]
-%if %3
-    mova      m7,  [srcq+mmsize*(4+%%i)]
-    mova      m8,  [srcq+mmsize*(5+%%i)]
-    mova      m9,  [srcq+mmsize*(6+%%i)]
-    mova      m10, [srcq+mmsize*(7+%%i)]
-%endif
-    CLIPD  m0,  m4, m5, m6
-    CLIPD  m1,  m4, m5, m6
-    CLIPD  m2,  m4, m5, m6
-    CLIPD  m3,  m4, m5, m6
-%if %3
-    CLIPD  m7,  m4, m5, m6
-    CLIPD  m8,  m4, m5, m6
-    CLIPD  m9,  m4, m5, m6
-    CLIPD  m10, m4, m5, m6
-%endif
-    mova  [dstq+mmsize*(0+%%i)], m0
-    mova  [dstq+mmsize*(1+%%i)], m1
-    mova  [dstq+mmsize*(2+%%i)], m2
-    mova  [dstq+mmsize*(3+%%i)], m3
-%if %3
-    mova  [dstq+mmsize*(4+%%i)], m7
-    mova  [dstq+mmsize*(5+%%i)], m8
-    mova  [dstq+mmsize*(6+%%i)], m9
-    mova  [dstq+mmsize*(7+%%i)], m10
-%endif
-%assign %%i %%i+4*(%3+1)
-%endrep
-    add     srcq, mmsize*4*(%2+%3)
-    add     dstq, mmsize*4*(%2+%3)
-    sub     lend, mmsize*(%2+%3)
-    jg .loop
-    REP_RET
-%endmacro
-
-INIT_MMX mmx
-%define CLIPD CLIPD_MMX
-VECTOR_CLIP_INT32 0, 1, 0, 0
-INIT_XMM sse2
-VECTOR_CLIP_INT32 6, 1, 0, 0, _int
-%define CLIPD CLIPD_SSE2
-VECTOR_CLIP_INT32 6, 2, 0, 1
-INIT_XMM sse4
-%define CLIPD CLIPD_SSE41
-%ifdef m8
-VECTOR_CLIP_INT32 11, 1, 1, 0
-%else
-VECTOR_CLIP_INT32 6, 1, 0, 0
-%endif
-
 ; %1 = aligned/unaligned
 %macro BSWAP_LOOPS  1
     mov      r3, r2
diff --git a/libavcodec/x86/dsputil_init.c b/libavcodec/x86/dsputil_init.c
index 5c12364c23..ed58598810 100644
--- a/libavcodec/x86/dsputil_init.c
+++ b/libavcodec/x86/dsputil_init.c
@@ -29,23 +29,9 @@
 #include "dsputil_x86.h"
 #include "idct_xvid.h"
 
-int32_t ff_scalarproduct_int16_mmxext(const int16_t *v1, const int16_t *v2,
-                                      int order);
-int32_t ff_scalarproduct_int16_sse2(const int16_t *v1, const int16_t *v2,
-                                    int order);
-
 void ff_bswap32_buf_ssse3(uint32_t *dst, const uint32_t *src, int w);
 void ff_bswap32_buf_sse2(uint32_t *dst, const uint32_t *src, int w);
 
-void ff_vector_clip_int32_mmx(int32_t *dst, const int32_t *src,
-                              int32_t min, int32_t max, unsigned int len);
-void ff_vector_clip_int32_sse2(int32_t *dst, const int32_t *src,
-                               int32_t min, int32_t max, unsigned int len);
-void ff_vector_clip_int32_int_sse2(int32_t *dst, const int32_t *src,
-                                   int32_t min, int32_t max, unsigned int len);
-void ff_vector_clip_int32_sse4(int32_t *dst, const int32_t *src,
-                               int32_t min, int32_t max, unsigned int len);
-
 static av_cold void dsputil_init_mmx(DSPContext *c, AVCodecContext *avctx,
                                      int cpu_flags, unsigned high_bit_depth)
 {
@@ -81,7 +67,6 @@ static av_cold void dsputil_init_mmx(DSPContext *c, AVCodecContext *avctx,
 #endif /* HAVE_MMX_INLINE */
 
 #if HAVE_MMX_EXTERNAL
-    c->vector_clip_int32 = ff_vector_clip_int32_mmx;
     c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_mmx;
 #endif /* HAVE_MMX_EXTERNAL */
 }
@@ -96,19 +81,12 @@ static av_cold void dsputil_init_mmxext(DSPContext *c, AVCodecContext *avctx,
         c->idct     = ff_idct_xvid_mmxext;
     }
 #endif /* HAVE_MMXEXT_INLINE */
-
-#if HAVE_MMXEXT_EXTERNAL
-    c->scalarproduct_int16          = ff_scalarproduct_int16_mmxext;
-#endif /* HAVE_MMXEXT_EXTERNAL */
 }
 
 static av_cold void dsputil_init_sse(DSPContext *c, AVCodecContext *avctx,
                                      int cpu_flags, unsigned high_bit_depth)
 {
 #if HAVE_YASM
-#if HAVE_SSE_EXTERNAL
-    c->vector_clipf = ff_vector_clipf_sse;
-#endif
 #if HAVE_INLINE_ASM && CONFIG_VIDEODSP
     c->gmc = ff_gmc_sse;
 #endif
@@ -128,12 +106,6 @@ static av_cold void dsputil_init_sse2(DSPContext *c, AVCodecContext *avctx,
 #endif /* HAVE_SSE2_INLINE */
 
 #if HAVE_SSE2_EXTERNAL
-    c->scalarproduct_int16          = ff_scalarproduct_int16_sse2;
-    if (cpu_flags & AV_CPU_FLAG_ATOM) {
-        c->vector_clip_int32 = ff_vector_clip_int32_int_sse2;
-    } else {
-        c->vector_clip_int32 = ff_vector_clip_int32_sse2;
-    }
     c->bswap_buf = ff_bswap32_buf_sse2;
     c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_sse2;
 #endif /* HAVE_SSE2_EXTERNAL */
@@ -147,14 +119,6 @@ static av_cold void dsputil_init_ssse3(DSPContext *c, AVCodecContext *avctx,
 #endif /* HAVE_SSSE3_EXTERNAL */
 }
 
-static av_cold void dsputil_init_sse4(DSPContext *c, AVCodecContext *avctx,
-                                      int cpu_flags, unsigned high_bit_depth)
-{
-#if HAVE_SSE4_EXTERNAL
-    c->vector_clip_int32 = ff_vector_clip_int32_sse4;
-#endif /* HAVE_SSE4_EXTERNAL */
-}
-
 av_cold void ff_dsputil_init_x86(DSPContext *c, AVCodecContext *avctx,
                                  unsigned high_bit_depth)
 {
@@ -175,9 +139,6 @@ av_cold void ff_dsputil_init_x86(DSPContext *c, AVCodecContext *avctx,
     if (EXTERNAL_SSSE3(cpu_flags))
         dsputil_init_ssse3(c, avctx, cpu_flags, high_bit_depth);
 
-    if (EXTERNAL_SSE4(cpu_flags))
-        dsputil_init_sse4(c, avctx, cpu_flags, high_bit_depth);
-
     if (CONFIG_ENCODERS)
         ff_dsputilenc_init_mmx(c, avctx, high_bit_depth);
 }
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c
index 3f187b70b5..54aba38b53 100644
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@@ -28,7 +28,6 @@
 #include "libavutil/x86/asm.h"
 #include "libavcodec/pixels.h"
 #include "libavcodec/videodsp.h"
-#include "constants.h"
 #include "dsputil_x86.h"
 #include "inline_asm.h"
 
diff --git a/libavcodec/x86/dsputil_x86.h b/libavcodec/x86/dsputil_x86.h
index e723df1937..b5d7291f28 100644
--- a/libavcodec/x86/dsputil_x86.h
+++ b/libavcodec/x86/dsputil_x86.h
@@ -53,10 +53,6 @@ void ff_gmc_sse(uint8_t *dst, uint8_t *src,
                 int dxx, int dxy, int dyx, int dyy,
                 int shift, int r, int width, int height);
 
-void ff_vector_clipf_sse(float *dst, const float *src,
-                         float min, float max, int len);
-
-
 void ff_mmx_idct(int16_t *block);
 void ff_mmxext_idct(int16_t *block);