Merge commit 'b2c087871dafc7d030b2d48457ddff597dfd4925'

* commit 'b2c087871dafc7d030b2d48457ddff597dfd4925':
  Move x86util.asm from libavcodec/ to libavutil/.
  Move x86inc.asm to libavutil/.
  APIchanges: note error_recognition in lavf
  lavf: add support for error_recognition, use it in avidec, and bump minor API version
  avconv: change semantics of -map
  avconv: get rid of new* options.
  cmdutils: allow precisely specifying a stream for AVOptions.
  configure: add missing CFLAGS to fix building on the HURD
  libx264: Include hint for possible values for configuring libx264
  cmdutils: allow ':'-separated modifiers in option names.
  avconv: make -map_metadata work consistently with the other options
  avconv: remove deprecated options.
  avconv: make -map_chapters accept only the input file index.
  Make a copy of ffmpeg under a new name -- avconv.
  ffmpeg: add a warning stating that the program is deprecated.
  Add weighted motion compensation for RV40 B-frames
  RV3/4: calculate B-frame motion weights once per frame
  Move RV3/4-specific DSP functions into their own context
  mjpeg: propagate decode errors from ff_mjpeg_decode_sos and ff_mjpeg_decode_dqt
  h264: notice memory allocation failure

Conflicts:
	.gitignore
	Makefile
	cmdutils.c
	configure
	doc/ffplay.texi
	doc/ffprobe.texi
	doc/ffserver.texi
	libavcodec/libx264.c
	libavformat/avformat.h
	libavformat/avidec.c
	libavformat/version.h
	tests/lavf-regression.sh
	tests/lavfi-regression.sh

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

27 files changed, 103 insertions, 1508 deletions

diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
index 09f65e8691..bce2c3ecaf 100644
--- a/libavcodec/x86/Makefile
+++ b/libavcodec/x86/Makefile
@@ -21,6 +21,8 @@ YASM-OBJS-$(CONFIG_H264PRED)           += x86/h264_intrapred.o          \
                                           x86/h264_intrapred_10bit.o
 MMX-OBJS-$(CONFIG_H264PRED)            += x86/h264_intrapred_init.o
 
+MMX-OBJS-$(CONFIG_RV40_DECODER)        += x86/rv40dsp.o                 \
+
 YASM-OBJS-$(CONFIG_VC1_DECODER)        += x86/vc1dsp_yasm.o
 
 MMX-OBJS-$(CONFIG_AC3DSP)              += x86/ac3dsp_mmx.o
diff --git a/libavcodec/x86/ac3dsp.asm b/libavcodec/x86/ac3dsp.asm
index 8c958a17ee..59157b7219 100644
--- a/libavcodec/x86/ac3dsp.asm
+++ b/libavcodec/x86/ac3dsp.asm
@@ -19,8 +19,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/dct32_sse.asm b/libavcodec/x86/dct32_sse.asm
index 9a2a6ea88c..bd69fe118c 100644
--- a/libavcodec/x86/dct32_sse.asm
+++ b/libavcodec/x86/dct32_sse.asm
@@ -19,8 +19,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA 32
 
diff --git a/libavcodec/x86/deinterlace.asm b/libavcodec/x86/deinterlace.asm
index 2997787beb..a09473bdae 100644
--- a/libavcodec/x86/deinterlace.asm
+++ b/libavcodec/x86/deinterlace.asm
@@ -20,8 +20,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c
index 5276053bb9..ea9d7f50ac 100644
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@@ -1799,29 +1799,17 @@ PREFETCH(prefetch_3dnow, prefetch)
 
 void ff_put_h264_chroma_mc8_mmx_rnd   (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_put_rv40_chroma_mc8_mmx       (uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
 void ff_avg_h264_chroma_mc8_mmx2_rnd  (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_avg_rv40_chroma_mc8_mmx2      (uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
 void ff_avg_h264_chroma_mc8_3dnow_rnd (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_avg_rv40_chroma_mc8_3dnow     (uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
 
 void ff_put_h264_chroma_mc4_mmx       (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_put_rv40_chroma_mc4_mmx       (uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
 void ff_avg_h264_chroma_mc4_mmx2      (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_avg_rv40_chroma_mc4_mmx2      (uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
 void ff_avg_h264_chroma_mc4_3dnow     (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_avg_rv40_chroma_mc4_3dnow     (uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
 
 void ff_put_h264_chroma_mc2_mmx2      (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
@@ -2476,9 +2464,6 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
         c->put_h264_chroma_pixels_tab[1]= ff_put_h264_chroma_mc4_mmx;
         }
 
-        c->put_rv40_chroma_pixels_tab[0]= ff_put_rv40_chroma_mc8_mmx;
-        c->put_rv40_chroma_pixels_tab[1]= ff_put_rv40_chroma_mc4_mmx;
-
         c->vector_clip_int32 = ff_vector_clip_int32_mmx;
 #endif
 
@@ -2578,9 +2563,6 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
             SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, mmx2, );
 
 #if HAVE_YASM
-            c->avg_rv40_chroma_pixels_tab[0]= ff_avg_rv40_chroma_mc8_mmx2;
-            c->avg_rv40_chroma_pixels_tab[1]= ff_avg_rv40_chroma_mc4_mmx2;
-
             if (!high_bit_depth) {
             c->avg_h264_chroma_pixels_tab[0]= ff_avg_h264_chroma_mc8_mmx2_rnd;
             c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_mmx2;
@@ -2662,8 +2644,6 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
             c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_3dnow;
             }
 
-            c->avg_rv40_chroma_pixels_tab[0]= ff_avg_rv40_chroma_mc8_3dnow;
-            c->avg_rv40_chroma_pixels_tab[1]= ff_avg_rv40_chroma_mc4_3dnow;
 #endif
         }
 
diff --git a/libavcodec/x86/dsputil_yasm.asm b/libavcodec/x86/dsputil_yasm.asm
index 1f5a4f68c7..7a646a2d36 100644
--- a/libavcodec/x86/dsputil_yasm.asm
+++ b/libavcodec/x86/dsputil_yasm.asm
@@ -19,7 +19,7 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
+%include "libavutil/x86/x86inc.asm"
 
 SECTION_RODATA
 pb_f: times 16 db 15
diff --git a/libavcodec/x86/dsputilenc_yasm.asm b/libavcodec/x86/dsputilenc_yasm.asm
index 016b354d6c..c08f53d39d 100644
--- a/libavcodec/x86/dsputilenc_yasm.asm
+++ b/libavcodec/x86/dsputilenc_yasm.asm
@@ -21,8 +21,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;*****************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION .text
 
diff --git a/libavcodec/x86/fft_mmx.asm b/libavcodec/x86/fft_mmx.asm
index 15fb67654c..8e5436770e 100644
--- a/libavcodec/x86/fft_mmx.asm
+++ b/libavcodec/x86/fft_mmx.asm
@@ -28,7 +28,7 @@
 ; in blocks as conventient to the vector size.
 ; i.e. {4x real, 4x imaginary, 4x real, ...} (or 2x respectively)
 
-%include "x86inc.asm"
+%include "libavutil/x86/x86inc.asm"
 
 %ifdef ARCH_X86_64
 %define pointer resq
diff --git a/libavcodec/x86/fmtconvert.asm b/libavcodec/x86/fmtconvert.asm
index 598e412f73..2deb577ca6 100644
--- a/libavcodec/x86/fmtconvert.asm
+++ b/libavcodec/x86/fmtconvert.asm
@@ -19,8 +19,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_TEXT
 
diff --git a/libavcodec/x86/h264_chromamc.asm b/libavcodec/x86/h264_chromamc.asm
index e9091f7059..72aecc0163 100644
--- a/libavcodec/x86/h264_chromamc.asm
+++ b/libavcodec/x86/h264_chromamc.asm
@@ -20,8 +20,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_chromamc_10bit.asm b/libavcodec/x86/h264_chromamc_10bit.asm
index 9d075434fe..56c0511857 100644
--- a/libavcodec/x86/h264_chromamc_10bit.asm
+++ b/libavcodec/x86/h264_chromamc_10bit.asm
@@ -22,8 +22,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_deblock.asm b/libavcodec/x86/h264_deblock.asm
index ca90f3f4c2..9595c0ab55 100644
--- a/libavcodec/x86/h264_deblock.asm
+++ b/libavcodec/x86/h264_deblock.asm
@@ -24,8 +24,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION .text
 
diff --git a/libavcodec/x86/h264_deblock_10bit.asm b/libavcodec/x86/h264_deblock_10bit.asm
index ee316258d3..b34867a36d 100644
--- a/libavcodec/x86/h264_deblock_10bit.asm
+++ b/libavcodec/x86/h264_deblock_10bit.asm
@@ -24,8 +24,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm
index 3b3dabe601..dd13bcd72f 100644
--- a/libavcodec/x86/h264_idct.asm
+++ b/libavcodec/x86/h264_idct.asm
@@ -26,8 +26,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;*****************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_idct_10bit.asm b/libavcodec/x86/h264_idct_10bit.asm
index 54636a95d0..a9c12da5a2 100644
--- a/libavcodec/x86/h264_idct_10bit.asm
+++ b/libavcodec/x86/h264_idct_10bit.asm
@@ -22,8 +22,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_intrapred.asm b/libavcodec/x86/h264_intrapred.asm
index a178a335ed..ed3bda7721 100644
--- a/libavcodec/x86/h264_intrapred.asm
+++ b/libavcodec/x86/h264_intrapred.asm
@@ -22,8 +22,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_intrapred_10bit.asm b/libavcodec/x86/h264_intrapred_10bit.asm
index e14e31a38c..5a06896afe 100644
--- a/libavcodec/x86/h264_intrapred_10bit.asm
+++ b/libavcodec/x86/h264_intrapred_10bit.asm
@@ -22,8 +22,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/h264_qpel_10bit.asm b/libavcodec/x86/h264_qpel_10bit.asm
index 15dd72ca36..30d0913cd1 100644
--- a/libavcodec/x86/h264_qpel_10bit.asm
+++ b/libavcodec/x86/h264_qpel_10bit.asm
@@ -22,8 +22,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA 32
 
diff --git a/libavcodec/x86/h264_weight.asm b/libavcodec/x86/h264_weight.asm
index 87544ae288..bb0af86097 100644
--- a/libavcodec/x86/h264_weight.asm
+++ b/libavcodec/x86/h264_weight.asm
@@ -21,7 +21,7 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
+%include "libavutil/x86/x86inc.asm"
 
 SECTION .text
 
diff --git a/libavcodec/x86/h264_weight_10bit.asm b/libavcodec/x86/h264_weight_10bit.asm
index 1c58d72d94..affc4ce5c4 100644
--- a/libavcodec/x86/h264_weight_10bit.asm
+++ b/libavcodec/x86/h264_weight_10bit.asm
@@ -22,8 +22,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA 32
 
diff --git a/libavcodec/x86/rv40dsp.c b/libavcodec/x86/rv40dsp.c
new file mode 100644
index 0000000000..9f90ad8bb6
--- /dev/null
+++ b/libavcodec/x86/rv40dsp.c
@@ -0,0 +1,60 @@
+/*
+ * RV40 decoder motion compensation functions x86-optimised
+ * Copyright (c) 2008 Konstantin Shishkov
+ *
+ * This file is part of Libav.
+ *
+ * Libav 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.
+ *
+ * Libav 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 Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * RV40 decoder motion compensation functions x86-optimised
+ */
+
+#include "libavcodec/rv34dsp.h"
+
+void ff_put_rv40_chroma_mc8_mmx  (uint8_t *dst, uint8_t *src,
+                                  int stride, int h, int x, int y);
+void ff_avg_rv40_chroma_mc8_mmx2 (uint8_t *dst, uint8_t *src,
+                                  int stride, int h, int x, int y);
+void ff_avg_rv40_chroma_mc8_3dnow(uint8_t *dst, uint8_t *src,
+                                  int stride, int h, int x, int y);
+
+void ff_put_rv40_chroma_mc4_mmx  (uint8_t *dst, uint8_t *src,
+                                  int stride, int h, int x, int y);
+void ff_avg_rv40_chroma_mc4_mmx2 (uint8_t *dst, uint8_t *src,
+                                  int stride, int h, int x, int y);
+void ff_avg_rv40_chroma_mc4_3dnow(uint8_t *dst, uint8_t *src,
+                                  int stride, int h, int x, int y);
+
+void ff_rv40dsp_init_x86(RV34DSPContext *c, DSPContext *dsp)
+{
+    av_unused int mm_flags = av_get_cpu_flags();
+
+#if HAVE_YASM
+    if (mm_flags & AV_CPU_FLAG_MMX) {
+        c->put_chroma_pixels_tab[0] = ff_put_rv40_chroma_mc8_mmx;
+        c->put_chroma_pixels_tab[1] = ff_put_rv40_chroma_mc4_mmx;
+    }
+    if (mm_flags & AV_CPU_FLAG_MMX2) {
+        c->avg_chroma_pixels_tab[0] = ff_avg_rv40_chroma_mc8_mmx2;
+        c->avg_chroma_pixels_tab[1] = ff_avg_rv40_chroma_mc4_mmx2;
+    } else if (mm_flags & AV_CPU_FLAG_3DNOW) {
+        c->avg_chroma_pixels_tab[0] = ff_avg_rv40_chroma_mc8_3dnow;
+        c->avg_chroma_pixels_tab[1] = ff_avg_rv40_chroma_mc4_3dnow;
+    }
+#endif
+}
diff --git a/libavcodec/x86/vc1dsp_yasm.asm b/libavcodec/x86/vc1dsp_yasm.asm
index a353c5f9ed..572aa0be1e 100644
--- a/libavcodec/x86/vc1dsp_yasm.asm
+++ b/libavcodec/x86/vc1dsp_yasm.asm
@@ -19,8 +19,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 cextern pw_4
 cextern pw_5
diff --git a/libavcodec/x86/vp3dsp.asm b/libavcodec/x86/vp3dsp.asm
index 328f94ca60..8dac6ab295 100644
--- a/libavcodec/x86/vp3dsp.asm
+++ b/libavcodec/x86/vp3dsp.asm
@@ -19,8 +19,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 ; MMX-optimized functions cribbed from the original VP3 source code.
 
diff --git a/libavcodec/x86/vp56dsp.asm b/libavcodec/x86/vp56dsp.asm
index 6da3ce30e3..98d8505c8a 100644
--- a/libavcodec/x86/vp56dsp.asm
+++ b/libavcodec/x86/vp56dsp.asm
@@ -20,8 +20,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 cextern pw_64
 
diff --git a/libavcodec/x86/vp8dsp.asm b/libavcodec/x86/vp8dsp.asm
index 14b49705dc..833c88a1a0 100644
--- a/libavcodec/x86/vp8dsp.asm
+++ b/libavcodec/x86/vp8dsp.asm
@@ -20,8 +20,8 @@
 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 ;******************************************************************************
 
-%include "x86inc.asm"
-%include "x86util.asm"
+%include "libavutil/x86/x86inc.asm"
+%include "libavutil/x86/x86util.asm"
 
 SECTION_RODATA
 
diff --git a/libavcodec/x86/x86inc.asm b/libavcodec/x86/x86inc.asm
deleted file mode 100644
index c84d5566a2..0000000000
--- a/libavcodec/x86/x86inc.asm
+++ /dev/null
@@ -1,905 +0,0 @@
-;*****************************************************************************
-;* x86inc.asm
-;*****************************************************************************
-;* Copyright (C) 2005-2011 x264 project
-;*
-;* Authors: Loren Merritt <lorenm@u.washington.edu>
-;*          Anton Mitrofanov <BugMaster@narod.ru>
-;*          Jason Garrett-Glaser <darkshikari@gmail.com>
-;*
-;* Permission to use, copy, modify, and/or distribute this software for any
-;* purpose with or without fee is hereby granted, provided that the above
-;* copyright notice and this permission notice appear in all copies.
-;*
-;* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
-;* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
-;* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
-;* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
-;* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
-;* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
-;* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
-;*****************************************************************************
-
-; This is a header file for the x264ASM assembly language, which uses
-; NASM/YASM syntax combined with a large number of macros to provide easy
-; abstraction between different calling conventions (x86_32, win64, linux64).
-; It also has various other useful features to simplify writing the kind of
-; DSP functions that are most often used in x264.
-
-; Unlike the rest of x264, this file is available under an ISC license, as it
-; has significant usefulness outside of x264 and we want it to be available
-; to the largest audience possible.  Of course, if you modify it for your own
-; purposes to add a new feature, we strongly encourage contributing a patch
-; as this feature might be useful for others as well.  Send patches or ideas
-; to x264-devel@videolan.org .
-
-%define program_name ff
-
-%ifdef ARCH_X86_64
-    %ifidn __OUTPUT_FORMAT__,win32
-        %define WIN64
-    %else
-        %define UNIX64
-    %endif
-%endif
-
-%ifdef PREFIX
-    %define mangle(x) _ %+ x
-%else
-    %define mangle(x) x
-%endif
-
-; FIXME: All of the 64bit asm functions that take a stride as an argument
-; via register, assume that the high dword of that register is filled with 0.
-; This is true in practice (since we never do any 64bit arithmetic on strides,
-; and x264's strides are all positive), but is not guaranteed by the ABI.
-
-; Name of the .rodata section.
-; Kludge: Something on OS X fails to align .rodata even given an align attribute,
-; so use a different read-only section.
-%macro SECTION_RODATA 0-1 16
-    %ifidn __OUTPUT_FORMAT__,macho64
-        SECTION .text align=%1
-    %elifidn __OUTPUT_FORMAT__,macho
-        SECTION .text align=%1
-        fakegot:
-    %elifidn __OUTPUT_FORMAT__,aout
-        section .text
-    %else
-        SECTION .rodata align=%1
-    %endif
-%endmacro
-
-; aout does not support align=
-%macro SECTION_TEXT 0-1 16
-    %ifidn __OUTPUT_FORMAT__,aout
-        SECTION .text
-    %else
-        SECTION .text align=%1
-    %endif
-%endmacro
-
-%ifdef WIN64
-    %define PIC
-%elifndef ARCH_X86_64
-; x86_32 doesn't require PIC.
-; Some distros prefer shared objects to be PIC, but nothing breaks if
-; the code contains a few textrels, so we'll skip that complexity.
-    %undef PIC
-%endif
-%ifdef PIC
-    default rel
-%endif
-
-; Macros to eliminate most code duplication between x86_32 and x86_64:
-; Currently this works only for leaf functions which load all their arguments
-; into registers at the start, and make no other use of the stack. Luckily that
-; covers most of x264's asm.
-
-; PROLOGUE:
-; %1 = number of arguments. loads them from stack if needed.
-; %2 = number of registers used. pushes callee-saved regs if needed.
-; %3 = number of xmm registers used. pushes callee-saved xmm regs if needed.
-; %4 = list of names to define to registers
-; PROLOGUE can also be invoked by adding the same options to cglobal
-
-; e.g.
-; cglobal foo, 2,3,0, dst, src, tmp
-; declares a function (foo), taking two args (dst and src) and one local variable (tmp)
-
-; TODO Some functions can use some args directly from the stack. If they're the
-; last args then you can just not declare them, but if they're in the middle
-; we need more flexible macro.
-
-; RET:
-; Pops anything that was pushed by PROLOGUE
-
-; REP_RET:
-; Same, but if it doesn't pop anything it becomes a 2-byte ret, for athlons
-; which are slow when a normal ret follows a branch.
-
-; registers:
-; rN and rNq are the native-size register holding function argument N
-; rNd, rNw, rNb are dword, word, and byte size
-; rNm is the original location of arg N (a register or on the stack), dword
-; rNmp is native size
-
-%macro DECLARE_REG 6
-    %define r%1q %2
-    %define r%1d %3
-    %define r%1w %4
-    %define r%1b %5
-    %define r%1m %6
-    %ifid %6 ; i.e. it's a register
-        %define r%1mp %2
-    %elifdef ARCH_X86_64 ; memory
-        %define r%1mp qword %6
-    %else
-        %define r%1mp dword %6
-    %endif
-    %define r%1  %2
-%endmacro
-
-%macro DECLARE_REG_SIZE 2
-    %define r%1q r%1
-    %define e%1q r%1
-    %define r%1d e%1
-    %define e%1d e%1
-    %define r%1w %1
-    %define e%1w %1
-    %define r%1b %2
-    %define e%1b %2
-%ifndef ARCH_X86_64
-    %define r%1  e%1
-%endif
-%endmacro
-
-DECLARE_REG_SIZE ax, al
-DECLARE_REG_SIZE bx, bl
-DECLARE_REG_SIZE cx, cl
-DECLARE_REG_SIZE dx, dl
-DECLARE_REG_SIZE si, sil
-DECLARE_REG_SIZE di, dil
-DECLARE_REG_SIZE bp, bpl
-
-; t# defines for when per-arch register allocation is more complex than just function arguments
-
-%macro DECLARE_REG_TMP 1-*
-    %assign %%i 0
-    %rep %0
-        CAT_XDEFINE t, %%i, r%1
-        %assign %%i %%i+1
-        %rotate 1
-    %endrep
-%endmacro
-
-%macro DECLARE_REG_TMP_SIZE 0-*
-    %rep %0
-        %define t%1q t%1 %+ q
-        %define t%1d t%1 %+ d
-        %define t%1w t%1 %+ w
-        %define t%1b t%1 %+ b
-        %rotate 1
-    %endrep
-%endmacro
-
-DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9
-
-%ifdef ARCH_X86_64
-    %define gprsize 8
-%else
-    %define gprsize 4
-%endif
-
-%macro PUSH 1
-    push %1
-    %assign stack_offset stack_offset+gprsize
-%endmacro
-
-%macro POP 1
-    pop %1
-    %assign stack_offset stack_offset-gprsize
-%endmacro
-
-%macro SUB 2
-    sub %1, %2
-    %ifidn %1, rsp
-        %assign stack_offset stack_offset+(%2)
-    %endif
-%endmacro
-
-%macro ADD 2
-    add %1, %2
-    %ifidn %1, rsp
-        %assign stack_offset stack_offset-(%2)
-    %endif
-%endmacro
-
-%macro movifnidn 2
-    %ifnidn %1, %2
-        mov %1, %2
-    %endif
-%endmacro
-
-%macro movsxdifnidn 2
-    %ifnidn %1, %2
-        movsxd %1, %2
-    %endif
-%endmacro
-
-%macro ASSERT 1
-    %if (%1) == 0
-        %error assert failed
-    %endif
-%endmacro
-
-%macro DEFINE_ARGS 0-*
-    %ifdef n_arg_names
-        %assign %%i 0
-        %rep n_arg_names
-            CAT_UNDEF arg_name %+ %%i, q
-            CAT_UNDEF arg_name %+ %%i, d
-            CAT_UNDEF arg_name %+ %%i, w
-            CAT_UNDEF arg_name %+ %%i, b
-            CAT_UNDEF arg_name %+ %%i, m
-            CAT_UNDEF arg_name, %%i
-            %assign %%i %%i+1
-        %endrep
-    %endif
-
-    %assign %%i 0
-    %rep %0
-        %xdefine %1q r %+ %%i %+ q
-        %xdefine %1d r %+ %%i %+ d
-        %xdefine %1w r %+ %%i %+ w
-        %xdefine %1b r %+ %%i %+ b
-        %xdefine %1m r %+ %%i %+ m
-        CAT_XDEFINE arg_name, %%i, %1
-        %assign %%i %%i+1
-        %rotate 1
-    %endrep
-    %assign n_arg_names %%i
-%endmacro
-
-%ifdef WIN64 ; Windows x64 ;=================================================
-
-DECLARE_REG 0, rcx, ecx, cx,  cl,  ecx
-DECLARE_REG 1, rdx, edx, dx,  dl,  edx
-DECLARE_REG 2, r8,  r8d, r8w, r8b, r8d
-DECLARE_REG 3, r9,  r9d, r9w, r9b, r9d
-DECLARE_REG 4, rdi, edi, di,  dil, [rsp + stack_offset + 40]
-DECLARE_REG 5, rsi, esi, si,  sil, [rsp + stack_offset + 48]
-DECLARE_REG 6, rax, eax, ax,  al,  [rsp + stack_offset + 56]
-%define r7m [rsp + stack_offset + 64]
-%define r8m [rsp + stack_offset + 72]
-
-%macro LOAD_IF_USED 2 ; reg_id, number_of_args
-    %if %1 < %2
-        mov r%1, [rsp + stack_offset + 8 + %1*8]
-    %endif
-%endmacro
-
-%macro PROLOGUE 2-4+ 0 ; #args, #regs, #xmm_regs, arg_names...
-    ASSERT %2 >= %1
-    %assign regs_used %2
-    ASSERT regs_used <= 7
-    %if regs_used > 4
-        push r4
-        push r5
-        %assign stack_offset stack_offset+16
-    %endif
-    WIN64_SPILL_XMM %3
-    LOAD_IF_USED 4, %1
-    LOAD_IF_USED 5, %1
-    LOAD_IF_USED 6, %1
-    DEFINE_ARGS %4
-%endmacro
-
-%macro WIN64_SPILL_XMM 1
-    %assign xmm_regs_used %1
-    ASSERT xmm_regs_used <= 16
-    %if xmm_regs_used > 6
-        sub rsp, (xmm_regs_used-6)*16+16
-        %assign stack_offset stack_offset+(xmm_regs_used-6)*16+16
-        %assign %%i xmm_regs_used
-        %rep (xmm_regs_used-6)
-            %assign %%i %%i-1
-            movdqa [rsp + (%%i-6)*16+8], xmm %+ %%i
-        %endrep
-    %endif
-%endmacro
-
-%macro WIN64_RESTORE_XMM_INTERNAL 1
-    %if xmm_regs_used > 6
-        %assign %%i xmm_regs_used
-        %rep (xmm_regs_used-6)
-            %assign %%i %%i-1
-            movdqa xmm %+ %%i, [%1 + (%%i-6)*16+8]
-        %endrep
-        add %1, (xmm_regs_used-6)*16+16
-    %endif
-%endmacro
-
-%macro WIN64_RESTORE_XMM 1
-    WIN64_RESTORE_XMM_INTERNAL %1
-    %assign stack_offset stack_offset-(xmm_regs_used-6)*16+16
-    %assign xmm_regs_used 0
-%endmacro
-
-%macro RET 0
-    WIN64_RESTORE_XMM_INTERNAL rsp
-    %if regs_used > 4
-        pop r5
-        pop r4
-    %endif
-    ret
-%endmacro
-
-%macro REP_RET 0
-    %if regs_used > 4 || xmm_regs_used > 6
-        RET
-    %else
-        rep ret
-    %endif
-%endmacro
-
-%elifdef ARCH_X86_64 ; *nix x64 ;=============================================
-
-DECLARE_REG 0, rdi, edi, di,  dil, edi
-DECLARE_REG 1, rsi, esi, si,  sil, esi
-DECLARE_REG 2, rdx, edx, dx,  dl,  edx
-DECLARE_REG 3, rcx, ecx, cx,  cl,  ecx
-DECLARE_REG 4, r8,  r8d, r8w, r8b, r8d
-DECLARE_REG 5, r9,  r9d, r9w, r9b, r9d
-DECLARE_REG 6, rax, eax, ax,  al,  [rsp + stack_offset + 8]
-%define r7m [rsp + stack_offset + 16]
-%define r8m [rsp + stack_offset + 24]
-
-%macro LOAD_IF_USED 2 ; reg_id, number_of_args
-    %if %1 < %2
-        mov r%1, [rsp - 40 + %1*8]
-    %endif
-%endmacro
-
-%macro PROLOGUE 2-4+ ; #args, #regs, #xmm_regs, arg_names...
-    ASSERT %2 >= %1
-    ASSERT %2 <= 7
-    LOAD_IF_USED 6, %1
-    DEFINE_ARGS %4
-%endmacro
-
-%macro RET 0
-    ret
-%endmacro
-
-%macro REP_RET 0
-    rep ret
-%endmacro
-
-%else ; X86_32 ;==============================================================
-
-DECLARE_REG 0, eax, eax, ax, al,   [esp + stack_offset + 4]
-DECLARE_REG 1, ecx, ecx, cx, cl,   [esp + stack_offset + 8]
-DECLARE_REG 2, edx, edx, dx, dl,   [esp + stack_offset + 12]
-DECLARE_REG 3, ebx, ebx, bx, bl,   [esp + stack_offset + 16]
-DECLARE_REG 4, esi, esi, si, null, [esp + stack_offset + 20]
-DECLARE_REG 5, edi, edi, di, null, [esp + stack_offset + 24]
-DECLARE_REG 6, ebp, ebp, bp, null, [esp + stack_offset + 28]
-%define r7m [esp + stack_offset + 32]
-%define r8m [esp + stack_offset + 36]
-%define rsp esp
-
-%macro PUSH_IF_USED 1 ; reg_id
-    %if %1 < regs_used
-        push r%1
-        %assign stack_offset stack_offset+4
-    %endif
-%endmacro
-
-%macro POP_IF_USED 1 ; reg_id
-    %if %1 < regs_used
-        pop r%1
-    %endif
-%endmacro
-
-%macro LOAD_IF_USED 2 ; reg_id, number_of_args
-    %if %1 < %2
-        mov r%1, [esp + stack_offset + 4 + %1*4]
-    %endif
-%endmacro
-
-%macro PROLOGUE 2-4+ ; #args, #regs, #xmm_regs, arg_names...
-    ASSERT %2 >= %1
-    %assign regs_used %2
-    ASSERT regs_used <= 7
-    PUSH_IF_USED 3
-    PUSH_IF_USED 4
-    PUSH_IF_USED 5
-    PUSH_IF_USED 6
-    LOAD_IF_USED 0, %1
-    LOAD_IF_USED 1, %1
-    LOAD_IF_USED 2, %1
-    LOAD_IF_USED 3, %1
-    LOAD_IF_USED 4, %1
-    LOAD_IF_USED 5, %1
-    LOAD_IF_USED 6, %1
-    DEFINE_ARGS %4
-%endmacro
-
-%macro RET 0
-    POP_IF_USED 6
-    POP_IF_USED 5
-    POP_IF_USED 4
-    POP_IF_USED 3
-    ret
-%endmacro
-
-%macro REP_RET 0
-    %if regs_used > 3
-        RET
-    %else
-        rep ret
-    %endif
-%endmacro
-
-%endif ;======================================================================
-
-%ifndef WIN64
-%macro WIN64_SPILL_XMM 1
-%endmacro
-%macro WIN64_RESTORE_XMM 1
-%endmacro
-%endif
-
-
-
-;=============================================================================
-; arch-independent part
-;=============================================================================
-
-%assign function_align 16
-
-; Symbol prefix for C linkage
-%macro cglobal 1-2+
-    %xdefine %1 mangle(program_name %+ _ %+ %1)
-    %xdefine %1.skip_prologue %1 %+ .skip_prologue
-    %ifidn __OUTPUT_FORMAT__,elf
-        global %1:function hidden
-    %else
-        global %1
-    %endif
-    align function_align
-    %1:
-    RESET_MM_PERMUTATION ; not really needed, but makes disassembly somewhat nicer
-    %assign stack_offset 0
-    %if %0 > 1
-        PROLOGUE %2
-    %endif
-%endmacro
-
-%macro cextern 1
-    %xdefine %1 mangle(program_name %+ _ %+ %1)
-    extern %1
-%endmacro
-
-;like cextern, but without the prefix
-%macro cextern_naked 1
-    %xdefine %1 mangle(%1)
-    extern %1
-%endmacro
-
-%macro const 2+
-    %xdefine %1 mangle(program_name %+ _ %+ %1)
-    global %1
-    %1: %2
-%endmacro
-
-; This is needed for ELF, otherwise the GNU linker assumes the stack is
-; executable by default.
-%ifidn __OUTPUT_FORMAT__,elf
-SECTION .note.GNU-stack noalloc noexec nowrite progbits
-%endif
-
-; merge mmx and sse*
-
-%macro CAT_XDEFINE 3
-    %xdefine %1%2 %3
-%endmacro
-
-%macro CAT_UNDEF 2
-    %undef %1%2
-%endmacro
-
-%macro INIT_MMX 0
-    %assign avx_enabled 0
-    %define RESET_MM_PERMUTATION INIT_MMX
-    %define mmsize 8
-    %define num_mmregs 8
-    %define mova movq
-    %define movu movq
-    %define movh movd
-    %define movnta movntq
-    %assign %%i 0
-    %rep 8
-    CAT_XDEFINE m, %%i, mm %+ %%i
-    CAT_XDEFINE nmm, %%i, %%i
-    %assign %%i %%i+1
-    %endrep
-    %rep 8
-    CAT_UNDEF m, %%i
-    CAT_UNDEF nmm, %%i
-    %assign %%i %%i+1
-    %endrep
-%endmacro
-
-%macro INIT_XMM 0
-    %assign avx_enabled 0
-    %define RESET_MM_PERMUTATION INIT_XMM
-    %define mmsize 16
-    %define num_mmregs 8
-    %ifdef ARCH_X86_64
-    %define num_mmregs 16
-    %endif
-    %define mova movdqa
-    %define movu movdqu
-    %define movh movq
-    %define movnta movntdq
-    %assign %%i 0
-    %rep num_mmregs
-    CAT_XDEFINE m, %%i, xmm %+ %%i
-    CAT_XDEFINE nxmm, %%i, %%i
-    %assign %%i %%i+1
-    %endrep
-%endmacro
-
-%macro INIT_AVX 0
-    INIT_XMM
-    %assign avx_enabled 1
-    %define PALIGNR PALIGNR_SSSE3
-    %define RESET_MM_PERMUTATION INIT_AVX
-%endmacro
-
-%macro INIT_YMM 0
-    %assign avx_enabled 1
-    %define RESET_MM_PERMUTATION INIT_YMM
-    %define mmsize 32
-    %define num_mmregs 8
-    %ifdef ARCH_X86_64
-    %define num_mmregs 16
-    %endif
-    %define mova vmovaps
-    %define movu vmovups
-    %assign %%i 0
-    %rep num_mmregs
-    CAT_XDEFINE m, %%i, ymm %+ %%i
-    CAT_XDEFINE nymm, %%i, %%i
-    %assign %%i %%i+1
-    %endrep
-%endmacro
-
-INIT_MMX
-
-; I often want to use macros that permute their arguments. e.g. there's no
-; efficient way to implement butterfly or transpose or dct without swapping some
-; arguments.
-;
-; I would like to not have to manually keep track of the permutations:
-; If I insert a permutation in the middle of a function, it should automatically
-; change everything that follows. For more complex macros I may also have multiple
-; implementations, e.g. the SSE2 and SSSE3 versions may have different permutations.
-;
-; Hence these macros. Insert a PERMUTE or some SWAPs at the end of a macro that
-; permutes its arguments. It's equivalent to exchanging the contents of the
-; registers, except that this way you exchange the register names instead, so it
-; doesn't cost any cycles.
-
-%macro PERMUTE 2-* ; takes a list of pairs to swap
-%rep %0/2
-    %xdefine tmp%2 m%2
-    %xdefine ntmp%2 nm%2
-    %rotate 2
-%endrep
-%rep %0/2
-    %xdefine m%1 tmp%2
-    %xdefine nm%1 ntmp%2
-    %undef tmp%2
-    %undef ntmp%2
-    %rotate 2
-%endrep
-%endmacro
-
-%macro SWAP 2-* ; swaps a single chain (sometimes more concise than pairs)
-%rep %0-1
-%ifdef m%1
-    %xdefine tmp m%1
-    %xdefine m%1 m%2
-    %xdefine m%2 tmp
-    CAT_XDEFINE n, m%1, %1
-    CAT_XDEFINE n, m%2, %2
-%else
-    ; If we were called as "SWAP m0,m1" rather than "SWAP 0,1" infer the original numbers here.
-    ; Be careful using this mode in nested macros though, as in some cases there may be
-    ; other copies of m# that have already been dereferenced and don't get updated correctly.
-    %xdefine %%n1 n %+ %1
-    %xdefine %%n2 n %+ %2
-    %xdefine tmp m %+ %%n1
-    CAT_XDEFINE m, %%n1, m %+ %%n2
-    CAT_XDEFINE m, %%n2, tmp
-    CAT_XDEFINE n, m %+ %%n1, %%n1
-    CAT_XDEFINE n, m %+ %%n2, %%n2
-%endif
-    %undef tmp
-    %rotate 1
-%endrep
-%endmacro
-
-; If SAVE_MM_PERMUTATION is placed at the end of a function and given the
-; function name, then any later calls to that function will automatically
-; load the permutation, so values can be returned in mmregs.
-%macro SAVE_MM_PERMUTATION 1 ; name to save as
-    %assign %%i 0
-    %rep num_mmregs
-    CAT_XDEFINE %1_m, %%i, m %+ %%i
-    %assign %%i %%i+1
-    %endrep
-%endmacro
-
-%macro LOAD_MM_PERMUTATION 1 ; name to load from
-    %assign %%i 0
-    %rep num_mmregs
-    CAT_XDEFINE m, %%i, %1_m %+ %%i
-    CAT_XDEFINE n, m %+ %%i, %%i
-    %assign %%i %%i+1
-    %endrep
-%endmacro
-
-%macro call 1
-    call %1
-    %ifdef %1_m0
-        LOAD_MM_PERMUTATION %1
-    %endif
-%endmacro
-
-; Substitutions that reduce instruction size but are functionally equivalent
-%macro add 2
-    %ifnum %2
-        %if %2==128
-            sub %1, -128
-        %else
-            add %1, %2
-        %endif
-    %else
-        add %1, %2
-    %endif
-%endmacro
-
-%macro sub 2
-    %ifnum %2
-        %if %2==128
-            add %1, -128
-        %else
-            sub %1, %2
-        %endif
-    %else
-        sub %1, %2
-    %endif
-%endmacro
-
-;=============================================================================
-; AVX abstraction layer
-;=============================================================================
-
-%assign i 0
-%rep 16
-    %if i < 8
-        CAT_XDEFINE sizeofmm, i, 8
-    %endif
-    CAT_XDEFINE sizeofxmm, i, 16
-    CAT_XDEFINE sizeofymm, i, 32
-%assign i i+1
-%endrep
-%undef i
-
-;%1 == instruction
-;%2 == 1 if float, 0 if int
-;%3 == 0 if 3-operand (xmm, xmm, xmm), 1 if 4-operand (xmm, xmm, xmm, imm)
-;%4 == number of operands given
-;%5+: operands
-%macro RUN_AVX_INSTR 6-7+
-    %if sizeof%5==32
-        v%1 %5, %6, %7
-    %else
-        %if sizeof%5==8
-            %define %%regmov movq
-        %elif %2
-            %define %%regmov movaps
-        %else
-            %define %%regmov movdqa
-        %endif
-
-        %if %4>=3+%3
-            %ifnidn %5, %6
-                %if avx_enabled && sizeof%5==16
-                    v%1 %5, %6, %7
-                %else
-                    %%regmov %5, %6
-                    %1 %5, %7
-                %endif
-            %else
-                %1 %5, %7
-            %endif
-        %elif %3
-            %1 %5, %6, %7
-        %else
-            %1 %5, %6
-        %endif
-    %endif
-%endmacro
-
-;%1 == instruction
-;%2 == 1 if float, 0 if int
-;%3 == 0 if 3-operand (xmm, xmm, xmm), 1 if 4-operand (xmm, xmm, xmm, imm)
-%macro AVX_INSTR 3
-    %macro %1 2-8 fnord, fnord, fnord, %1, %2, %3
-        %ifidn %3, fnord
-            RUN_AVX_INSTR %6, %7, %8, 2, %1, %2
-        %elifidn %4, fnord
-            RUN_AVX_INSTR %6, %7, %8, 3, %1, %2, %3
-        %elifidn %5, fnord
-            RUN_AVX_INSTR %6, %7, %8, 4, %1, %2, %3, %4
-        %else
-            RUN_AVX_INSTR %6, %7, %8, 5, %1, %2, %3, %4, %5
-        %endif
-    %endmacro
-%endmacro
-
-AVX_INSTR addpd, 1, 0
-AVX_INSTR addps, 1, 0
-AVX_INSTR addsd, 1, 0
-AVX_INSTR addss, 1, 0
-AVX_INSTR addsubpd, 1, 0
-AVX_INSTR addsubps, 1, 0
-AVX_INSTR andpd, 1, 0
-AVX_INSTR andps, 1, 0
-AVX_INSTR andnpd, 1, 0
-AVX_INSTR andnps, 1, 0
-AVX_INSTR blendpd, 1, 0
-AVX_INSTR blendps, 1, 0
-AVX_INSTR blendvpd, 1, 0
-AVX_INSTR blendvps, 1, 0
-AVX_INSTR cmppd, 1, 0
-AVX_INSTR cmpps, 1, 0
-AVX_INSTR cmpsd, 1, 0
-AVX_INSTR cmpss, 1, 0
-AVX_INSTR divpd, 1, 0
-AVX_INSTR divps, 1, 0
-AVX_INSTR divsd, 1, 0
-AVX_INSTR divss, 1, 0
-AVX_INSTR dppd, 1, 0
-AVX_INSTR dpps, 1, 0
-AVX_INSTR haddpd, 1, 0
-AVX_INSTR haddps, 1, 0
-AVX_INSTR hsubpd, 1, 0
-AVX_INSTR hsubps, 1, 0
-AVX_INSTR maxpd, 1, 0
-AVX_INSTR maxps, 1, 0
-AVX_INSTR maxsd, 1, 0
-AVX_INSTR maxss, 1, 0
-AVX_INSTR minpd, 1, 0
-AVX_INSTR minps, 1, 0
-AVX_INSTR minsd, 1, 0
-AVX_INSTR minss, 1, 0
-AVX_INSTR mpsadbw, 0, 1
-AVX_INSTR mulpd, 1, 0
-AVX_INSTR mulps, 1, 0
-AVX_INSTR mulsd, 1, 0
-AVX_INSTR mulss, 1, 0
-AVX_INSTR orpd, 1, 0
-AVX_INSTR orps, 1, 0
-AVX_INSTR packsswb, 0, 0
-AVX_INSTR packssdw, 0, 0
-AVX_INSTR packuswb, 0, 0
-AVX_INSTR packusdw, 0, 0
-AVX_INSTR paddb, 0, 0
-AVX_INSTR paddw, 0, 0
-AVX_INSTR paddd, 0, 0
-AVX_INSTR paddq, 0, 0
-AVX_INSTR paddsb, 0, 0
-AVX_INSTR paddsw, 0, 0
-AVX_INSTR paddusb, 0, 0
-AVX_INSTR paddusw, 0, 0
-AVX_INSTR palignr, 0, 1
-AVX_INSTR pand, 0, 0
-AVX_INSTR pandn, 0, 0
-AVX_INSTR pavgb, 0, 0
-AVX_INSTR pavgw, 0, 0
-AVX_INSTR pblendvb, 0, 0
-AVX_INSTR pblendw, 0, 1
-AVX_INSTR pcmpestri, 0, 0
-AVX_INSTR pcmpestrm, 0, 0
-AVX_INSTR pcmpistri, 0, 0
-AVX_INSTR pcmpistrm, 0, 0
-AVX_INSTR pcmpeqb, 0, 0
-AVX_INSTR pcmpeqw, 0, 0
-AVX_INSTR pcmpeqd, 0, 0
-AVX_INSTR pcmpeqq, 0, 0
-AVX_INSTR pcmpgtb, 0, 0
-AVX_INSTR pcmpgtw, 0, 0
-AVX_INSTR pcmpgtd, 0, 0
-AVX_INSTR pcmpgtq, 0, 0
-AVX_INSTR phaddw, 0, 0
-AVX_INSTR phaddd, 0, 0
-AVX_INSTR phaddsw, 0, 0
-AVX_INSTR phsubw, 0, 0
-AVX_INSTR phsubd, 0, 0
-AVX_INSTR phsubsw, 0, 0
-AVX_INSTR pmaddwd, 0, 0
-AVX_INSTR pmaddubsw, 0, 0
-AVX_INSTR pmaxsb, 0, 0
-AVX_INSTR pmaxsw, 0, 0
-AVX_INSTR pmaxsd, 0, 0
-AVX_INSTR pmaxub, 0, 0
-AVX_INSTR pmaxuw, 0, 0
-AVX_INSTR pmaxud, 0, 0
-AVX_INSTR pminsb, 0, 0
-AVX_INSTR pminsw, 0, 0
-AVX_INSTR pminsd, 0, 0
-AVX_INSTR pminub, 0, 0
-AVX_INSTR pminuw, 0, 0
-AVX_INSTR pminud, 0, 0
-AVX_INSTR pmulhuw, 0, 0
-AVX_INSTR pmulhrsw, 0, 0
-AVX_INSTR pmulhw, 0, 0
-AVX_INSTR pmullw, 0, 0
-AVX_INSTR pmulld, 0, 0
-AVX_INSTR pmuludq, 0, 0
-AVX_INSTR pmuldq, 0, 0
-AVX_INSTR por, 0, 0
-AVX_INSTR psadbw, 0, 0
-AVX_INSTR pshufb, 0, 0
-AVX_INSTR psignb, 0, 0
-AVX_INSTR psignw, 0, 0
-AVX_INSTR psignd, 0, 0
-AVX_INSTR psllw, 0, 0
-AVX_INSTR pslld, 0, 0
-AVX_INSTR psllq, 0, 0
-AVX_INSTR pslldq, 0, 0
-AVX_INSTR psraw, 0, 0
-AVX_INSTR psrad, 0, 0
-AVX_INSTR psrlw, 0, 0
-AVX_INSTR psrld, 0, 0
-AVX_INSTR psrlq, 0, 0
-AVX_INSTR psrldq, 0, 0
-AVX_INSTR psubb, 0, 0
-AVX_INSTR psubw, 0, 0
-AVX_INSTR psubd, 0, 0
-AVX_INSTR psubq, 0, 0
-AVX_INSTR psubsb, 0, 0
-AVX_INSTR psubsw, 0, 0
-AVX_INSTR psubusb, 0, 0
-AVX_INSTR psubusw, 0, 0
-AVX_INSTR punpckhbw, 0, 0
-AVX_INSTR punpckhwd, 0, 0
-AVX_INSTR punpckhdq, 0, 0
-AVX_INSTR punpckhqdq, 0, 0
-AVX_INSTR punpcklbw, 0, 0
-AVX_INSTR punpcklwd, 0, 0
-AVX_INSTR punpckldq, 0, 0
-AVX_INSTR punpcklqdq, 0, 0
-AVX_INSTR pxor, 0, 0
-AVX_INSTR shufps, 0, 1
-AVX_INSTR subpd, 1, 0
-AVX_INSTR subps, 1, 0
-AVX_INSTR subsd, 1, 0
-AVX_INSTR subss, 1, 0
-AVX_INSTR unpckhpd, 1, 0
-AVX_INSTR unpckhps, 1, 0
-AVX_INSTR unpcklpd, 1, 0
-AVX_INSTR unpcklps, 1, 0
-AVX_INSTR xorpd, 1, 0
-AVX_INSTR xorps, 1, 0
-
-; 3DNow instructions, for sharing code between AVX, SSE and 3DN
-AVX_INSTR pfadd, 1, 0
-AVX_INSTR pfsub, 1, 0
-AVX_INSTR pfmul, 1, 0
diff --git a/libavcodec/x86/x86util.asm b/libavcodec/x86/x86util.asm
deleted file mode 100644
index 1cede4d336..0000000000
--- a/libavcodec/x86/x86util.asm
+++ /dev/null
@@ -1,542 +0,0 @@
-;*****************************************************************************
-;* x86util.asm
-;*****************************************************************************
-;* Copyright (C) 2008-2010 x264 project
-;*
-;* Authors: Loren Merritt <lorenm@u.washington.edu>
-;*          Holger Lubitz <holger@lubitz.org>
-;*
-;* 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
-;******************************************************************************
-
-%macro SBUTTERFLY 4
-%if avx_enabled == 0
-    mova      m%4, m%2
-    punpckl%1 m%2, m%3
-    punpckh%1 m%4, m%3
-%else
-    punpckh%1 m%4, m%2, m%3
-    punpckl%1 m%2, m%3
-%endif
-    SWAP %3, %4
-%endmacro
-
-%macro SBUTTERFLY2 4
-    punpckl%1 m%4, m%2, m%3
-    punpckh%1 m%2, m%2, m%3
-    SWAP %2, %4, %3
-%endmacro
-
-%macro SBUTTERFLYPS 3
-    movaps   m%3, m%1
-    unpcklps m%1, m%2
-    unpckhps m%3, m%2
-    SWAP %2, %3
-%endmacro
-
-%macro TRANSPOSE4x4B 5
-    SBUTTERFLY bw, %1, %2, %5
-    SBUTTERFLY bw, %3, %4, %5
-    SBUTTERFLY wd, %1, %3, %5
-    SBUTTERFLY wd, %2, %4, %5
-    SWAP %2, %3
-%endmacro
-
-%macro TRANSPOSE4x4W 5
-    SBUTTERFLY wd, %1, %2, %5
-    SBUTTERFLY wd, %3, %4, %5
-    SBUTTERFLY dq, %1, %3, %5
-    SBUTTERFLY dq, %2, %4, %5
-    SWAP %2, %3
-%endmacro
-
-%macro TRANSPOSE2x4x4W 5
-    SBUTTERFLY wd,  %1, %2, %5
-    SBUTTERFLY wd,  %3, %4, %5
-    SBUTTERFLY dq,  %1, %3, %5
-    SBUTTERFLY dq,  %2, %4, %5
-    SBUTTERFLY qdq, %1, %2, %5
-    SBUTTERFLY qdq, %3, %4, %5
-%endmacro
-
-%macro TRANSPOSE4x4D 5
-    SBUTTERFLY dq,  %1, %2, %5
-    SBUTTERFLY dq,  %3, %4, %5
-    SBUTTERFLY qdq, %1, %3, %5
-    SBUTTERFLY qdq, %2, %4, %5
-    SWAP %2, %3
-%endmacro
-
-; identical behavior to TRANSPOSE4x4D, but using SSE1 float ops
-%macro TRANSPOSE4x4PS 5
-    SBUTTERFLYPS %1, %2, %5
-    SBUTTERFLYPS %3, %4, %5
-    movaps  m%5, m%1
-    movlhps m%1, m%3
-    movhlps m%3, m%5
-    movaps  m%5, m%2
-    movlhps m%2, m%4
-    movhlps m%4, m%5
-    SWAP %2, %3
-%endmacro
-
-%macro TRANSPOSE8x8W 9-11
-%ifdef ARCH_X86_64
-    SBUTTERFLY wd,  %1, %2, %9
-    SBUTTERFLY wd,  %3, %4, %9
-    SBUTTERFLY wd,  %5, %6, %9
-    SBUTTERFLY wd,  %7, %8, %9
-    SBUTTERFLY dq,  %1, %3, %9
-    SBUTTERFLY dq,  %2, %4, %9
-    SBUTTERFLY dq,  %5, %7, %9
-    SBUTTERFLY dq,  %6, %8, %9
-    SBUTTERFLY qdq, %1, %5, %9
-    SBUTTERFLY qdq, %2, %6, %9
-    SBUTTERFLY qdq, %3, %7, %9
-    SBUTTERFLY qdq, %4, %8, %9
-    SWAP %2, %5
-    SWAP %4, %7
-%else
-; in:  m0..m7, unless %11 in which case m6 is in %9
-; out: m0..m7, unless %11 in which case m4 is in %10
-; spills into %9 and %10
-%if %0<11
-    movdqa %9, m%7
-%endif
-    SBUTTERFLY wd,  %1, %2, %7
-    movdqa %10, m%2
-    movdqa m%7, %9
-    SBUTTERFLY wd,  %3, %4, %2
-    SBUTTERFLY wd,  %5, %6, %2
-    SBUTTERFLY wd,  %7, %8, %2
-    SBUTTERFLY dq,  %1, %3, %2
-    movdqa %9, m%3
-    movdqa m%2, %10
-    SBUTTERFLY dq,  %2, %4, %3
-    SBUTTERFLY dq,  %5, %7, %3
-    SBUTTERFLY dq,  %6, %8, %3
-    SBUTTERFLY qdq, %1, %5, %3
-    SBUTTERFLY qdq, %2, %6, %3
-    movdqa %10, m%2
-    movdqa m%3, %9
-    SBUTTERFLY qdq, %3, %7, %2
-    SBUTTERFLY qdq, %4, %8, %2
-    SWAP %2, %5
-    SWAP %4, %7
-%if %0<11
-    movdqa m%5, %10
-%endif
-%endif
-%endmacro
-
-; PABSW macros assume %1 != %2, while ABS1/2 macros work in-place
-%macro PABSW_MMX 2
-    pxor       %1, %1
-    pcmpgtw    %1, %2
-    pxor       %2, %1
-    psubw      %2, %1
-    SWAP       %1, %2
-%endmacro
-
-%macro PSIGNW_MMX 2
-    pxor       %1, %2
-    psubw      %1, %2
-%endmacro
-
-%macro PABSW_MMX2 2
-    pxor    %1, %1
-    psubw   %1, %2
-    pmaxsw  %1, %2
-%endmacro
-
-%macro PABSW_SSSE3 2
-    pabsw      %1, %2
-%endmacro
-
-%macro PSIGNW_SSSE3 2
-    psignw     %1, %2
-%endmacro
-
-%macro ABS1_MMX 2    ; a, tmp
-    pxor       %2, %2
-    pcmpgtw    %2, %1
-    pxor       %1, %2
-    psubw      %1, %2
-%endmacro
-
-%macro ABS2_MMX 4    ; a, b, tmp0, tmp1
-    pxor       %3, %3
-    pxor       %4, %4
-    pcmpgtw    %3, %1
-    pcmpgtw    %4, %2
-    pxor       %1, %3
-    pxor       %2, %4
-    psubw      %1, %3
-    psubw      %2, %4
-%endmacro
-
-%macro ABS1_MMX2 2   ; a, tmp
-    pxor    %2, %2
-    psubw   %2, %1
-    pmaxsw  %1, %2
-%endmacro
-
-%macro ABS2_MMX2 4   ; a, b, tmp0, tmp1
-    pxor    %3, %3
-    pxor    %4, %4
-    psubw   %3, %1
-    psubw   %4, %2
-    pmaxsw  %1, %3
-    pmaxsw  %2, %4
-%endmacro
-
-%macro ABS1_SSSE3 2
-    pabsw   %1, %1
-%endmacro
-
-%macro ABS2_SSSE3 4
-    pabsw   %1, %1
-    pabsw   %2, %2
-%endmacro
-
-%macro ABSB_MMX 2
-    pxor    %2, %2
-    psubb   %2, %1
-    pminub  %1, %2
-%endmacro
-
-%macro ABSB2_MMX 4
-    pxor    %3, %3
-    pxor    %4, %4
-    psubb   %3, %1
-    psubb   %4, %2
-    pminub  %1, %3
-    pminub  %2, %4
-%endmacro
-
-%macro ABSD2_MMX 4
-    pxor    %3, %3
-    pxor    %4, %4
-    pcmpgtd %3, %1
-    pcmpgtd %4, %2
-    pxor    %1, %3
-    pxor    %2, %4
-    psubd   %1, %3
-    psubd   %2, %4
-%endmacro
-
-%macro ABSB_SSSE3 2
-    pabsb   %1, %1
-%endmacro
-
-%macro ABSB2_SSSE3 4
-    pabsb   %1, %1
-    pabsb   %2, %2
-%endmacro
-
-%macro ABS4 6
-    ABS2 %1, %2, %5, %6
-    ABS2 %3, %4, %5, %6
-%endmacro
-
-%define ABS1 ABS1_MMX
-%define ABS2 ABS2_MMX
-%define ABSB ABSB_MMX
-%define ABSB2 ABSB2_MMX
-
-%macro SPLATB_MMX 3
-    movd      %1, [%2-3] ;to avoid crossing a cacheline
-    punpcklbw %1, %1
-    SPLATW    %1, %1, 3
-%endmacro
-
-%macro SPLATB_SSSE3 3
-    movd      %1, [%2-3]
-    pshufb    %1, %3
-%endmacro
-
-%macro PALIGNR_MMX 4-5 ; [dst,] src1, src2, imm, tmp
-    %define %%dst %1
-%if %0==5
-%ifnidn %1, %2
-    mova    %%dst, %2
-%endif
-    %rotate 1
-%endif
-%ifnidn %4, %2
-    mova    %4, %2
-%endif
-%if mmsize==8
-    psllq   %%dst, (8-%3)*8
-    psrlq   %4, %3*8
-%else
-    pslldq  %%dst, 16-%3
-    psrldq  %4, %3
-%endif
-    por     %%dst, %4
-%endmacro
-
-%macro PALIGNR_SSSE3 4-5
-%if %0==5
-    palignr %1, %2, %3, %4
-%else
-    palignr %1, %2, %3
-%endif
-%endmacro
-
-%macro DEINTB 5 ; mask, reg1, mask, reg2, optional src to fill masks from
-%ifnum %5
-    pand   m%3, m%5, m%4 ; src .. y6 .. y4
-    pand   m%1, m%5, m%2 ; dst .. y6 .. y4
-%else
-    mova   m%1, %5
-    pand   m%3, m%1, m%4 ; src .. y6 .. y4
-    pand   m%1, m%1, m%2 ; dst .. y6 .. y4
-%endif
-    psrlw  m%2, 8        ; dst .. y7 .. y5
-    psrlw  m%4, 8        ; src .. y7 .. y5
-%endmacro
-
-%macro SUMSUB_BA 3-4
-%if %0==3
-    padd%1  m%2, m%3
-    padd%1  m%3, m%3
-    psub%1  m%3, m%2
-%else
-%if avx_enabled == 0
-    mova    m%4, m%2
-    padd%1  m%2, m%3
-    psub%1  m%3, m%4
-%else
-    padd%1  m%4, m%2, m%3
-    psub%1  m%3, m%2
-    SWAP    %2, %4
-%endif
-%endif
-%endmacro
-
-%macro SUMSUB_BADC 5-6
-%if %0==6
-    SUMSUB_BA %1, %2, %3, %6
-    SUMSUB_BA %1, %4, %5, %6
-%else
-    padd%1  m%2, m%3
-    padd%1  m%4, m%5
-    padd%1  m%3, m%3
-    padd%1  m%5, m%5
-    psub%1  m%3, m%2
-    psub%1  m%5, m%4
-%endif
-%endmacro
-
-%macro SUMSUB2_AB 4
-%ifnum %3
-    psub%1  m%4, m%2, m%3
-    psub%1  m%4, m%3
-    padd%1  m%2, m%2
-    padd%1  m%2, m%3
-%else
-    mova    m%4, m%2
-    padd%1  m%2, m%2
-    padd%1  m%2, %3
-    psub%1  m%4, %3
-    psub%1  m%4, %3
-%endif
-%endmacro
-
-%macro SUMSUB2_BA 4
-%if avx_enabled == 0
-    mova    m%4, m%2
-    padd%1  m%2, m%3
-    padd%1  m%2, m%3
-    psub%1  m%3, m%4
-    psub%1  m%3, m%4
-%else
-    padd%1  m%4, m%2, m%3
-    padd%1  m%4, m%3
-    psub%1  m%3, m%2
-    psub%1  m%3, m%2
-    SWAP     %2,  %4
-%endif
-%endmacro
-
-%macro SUMSUBD2_AB 5
-%ifnum %4
-    psra%1  m%5, m%2, 1  ; %3: %3>>1
-    psra%1  m%4, m%3, 1  ; %2: %2>>1
-    padd%1  m%4, m%2     ; %3: %3>>1+%2
-    psub%1  m%5, m%3     ; %2: %2>>1-%3
-    SWAP     %2, %5
-    SWAP     %3, %4
-%else
-    mova    %5, m%2
-    mova    %4, m%3
-    psra%1  m%3, 1  ; %3: %3>>1
-    psra%1  m%2, 1  ; %2: %2>>1
-    padd%1  m%3, %5 ; %3: %3>>1+%2
-    psub%1  m%2, %4 ; %2: %2>>1-%3
-%endif
-%endmacro
-
-%macro DCT4_1D 5
-%ifnum %5
-    SUMSUB_BADC w, %4, %1, %3, %2, %5
-    SUMSUB_BA   w, %3, %4, %5
-    SUMSUB2_AB  w, %1, %2, %5
-    SWAP %1, %3, %4, %5, %2
-%else
-    SUMSUB_BADC w, %4, %1, %3, %2
-    SUMSUB_BA   w, %3, %4
-    mova     [%5], m%2
-    SUMSUB2_AB  w, %1, [%5], %2
-    SWAP %1, %3, %4, %2
-%endif
-%endmacro
-
-%macro IDCT4_1D 6-7
-%ifnum %6
-    SUMSUBD2_AB %1, %3, %5, %7, %6
-    ; %3: %3>>1-%5 %5: %3+%5>>1
-    SUMSUB_BA   %1, %4, %2, %7
-    ; %4: %2+%4 %2: %2-%4
-    SUMSUB_BADC %1, %5, %4, %3, %2, %7
-    ; %5: %2+%4 + (%3+%5>>1)
-    ; %4: %2+%4 - (%3+%5>>1)
-    ; %3: %2-%4 + (%3>>1-%5)
-    ; %2: %2-%4 - (%3>>1-%5)
-%else
-%ifidn %1, w
-    SUMSUBD2_AB %1, %3, %5, [%6], [%6+16]
-%else
-    SUMSUBD2_AB %1, %3, %5, [%6], [%6+32]
-%endif
-    SUMSUB_BA   %1, %4, %2
-    SUMSUB_BADC %1, %5, %4, %3, %2
-%endif
-    SWAP %2, %5, %4
-    ; %2: %2+%4 + (%3+%5>>1) row0
-    ; %3: %2-%4 + (%3>>1-%5) row1
-    ; %4: %2-%4 - (%3>>1-%5) row2
-    ; %5: %2+%4 - (%3+%5>>1) row3
-%endmacro
-
-
-%macro LOAD_DIFF 5
-%ifidn %3, none
-    movh       %1, %4
-    movh       %2, %5
-    punpcklbw  %1, %2
-    punpcklbw  %2, %2
-    psubw      %1, %2
-%else
-    movh       %1, %4
-    punpcklbw  %1, %3
-    movh       %2, %5
-    punpcklbw  %2, %3
-    psubw      %1, %2
-%endif
-%endmacro
-
-%macro STORE_DCT 6
-    movq   [%5+%6+ 0], m%1
-    movq   [%5+%6+ 8], m%2
-    movq   [%5+%6+16], m%3
-    movq   [%5+%6+24], m%4
-    movhps [%5+%6+32], m%1
-    movhps [%5+%6+40], m%2
-    movhps [%5+%6+48], m%3
-    movhps [%5+%6+56], m%4
-%endmacro
-
-%macro LOAD_DIFF_8x4P 7-10 r0,r2,0 ; 4x dest, 2x temp, 2x pointer, increment?
-    LOAD_DIFF m%1, m%5, m%7, [%8],      [%9]
-    LOAD_DIFF m%2, m%6, m%7, [%8+r1],   [%9+r3]
-    LOAD_DIFF m%3, m%5, m%7, [%8+2*r1], [%9+2*r3]
-    LOAD_DIFF m%4, m%6, m%7, [%8+r4],   [%9+r5]
-%if %10
-    lea %8, [%8+4*r1]
-    lea %9, [%9+4*r3]
-%endif
-%endmacro
-
-%macro DIFFx2 6-7
-    movh       %3, %5
-    punpcklbw  %3, %4
-    psraw      %1, 6
-    paddsw     %1, %3
-    movh       %3, %6
-    punpcklbw  %3, %4
-    psraw      %2, 6
-    paddsw     %2, %3
-    packuswb   %2, %1
-%endmacro
-
-%macro STORE_DIFF 4
-    movh       %2, %4
-    punpcklbw  %2, %3
-    psraw      %1, 6
-    paddsw     %1, %2
-    packuswb   %1, %1
-    movh       %4, %1
-%endmacro
-
-%macro STORE_DIFFx2 8 ; add1, add2, reg1, reg2, zero, shift, source, stride
-    movh       %3, [%7]
-    movh       %4, [%7+%8]
-    psraw      %1, %6
-    psraw      %2, %6
-    punpcklbw  %3, %5
-    punpcklbw  %4, %5
-    paddw      %3, %1
-    paddw      %4, %2
-    packuswb   %3, %5
-    packuswb   %4, %5
-    movh     [%7], %3
-    movh  [%7+%8], %4
-%endmacro
-
-%macro PMINUB_MMX 3 ; dst, src, tmp
-    mova     %3, %1
-    psubusb  %3, %2
-    psubb    %1, %3
-%endmacro
-
-%macro PMINUB_MMXEXT 3 ; dst, src, ignored
-    pminub   %1, %2
-%endmacro
-
-%macro SPLATW 2-3 0
-%if mmsize == 16
-    pshuflw    %1, %2, (%3)*0x55
-    punpcklqdq %1, %1
-%else
-    pshufw     %1, %2, (%3)*0x55
-%endif
-%endmacro
-
-%macro SPLATD 2-3 0
-%if mmsize == 16
-    pshufd %1, %2, (%3)*0x55
-%else
-    pshufw %1, %2, (%3)*0x11 + ((%3)+1)*0x44
-%endif
-%endmacro
-
-%macro CLIPW 3 ;(dst, min, max)
-    pmaxsw %1, %2
-    pminsw %1, %3
-%endmacro