Merge commit 'c65d67ef'

* commit 'c65d67ef':
  aarch64: hpeldsp NEON optimizations

Conflicts:
	libavcodec/hpeldsp.c
	libavcodec/hpeldsp.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/aarch64/Makefile

 OBJS-$(CONFIG_H264CHROMA)               += aarch64/h264chroma_init_aarch64.o
 OBJS-$(CONFIG_H264DSP)                  += aarch64/h264dsp_init_aarch64.o
 OBJS-$(CONFIG_H264QPEL)                 += aarch64/h264qpel_init_aarch64.o
+OBJS-$(CONFIG_HPELDSP)                  += aarch64/hpeldsp_init_aarch64.o
 OBJS-$(CONFIG_RV40_DECODER)             += aarch64/rv40dsp_init_aarch64.o
 OBJS-$(CONFIG_VC1_DECODER)              += aarch64/vc1dsp_init_aarch64.o
 
 NEON-OBJS-$(CONFIG_H264CHROMA)          += aarch64/h264cmc_neon.o
 NEON-OBJS-$(CONFIG_H264DSP)             += aarch64/h264idct_neon.o
-NEON-OBJS-$(CONFIG_H264QPEL)            += aarch64/h264qpel_neon.o
+NEON-OBJS-$(CONFIG_H264QPEL)            += aarch64/h264qpel_neon.o             \
+                                           aarch64/hpeldsp_neon.o
+NEON-OBJS-$(CONFIG_HPELDSP)             += aarch64/hpeldsp_neon.o

libavcodec/aarch64/h264qpel_init_aarch64.c

@@ -101,7 +101,7 @@ av_cold void ff_h264qpel_init_aarch64(H264QpelContext *c, int bit_depth)
     int cpu_flags = av_get_cpu_flags();
 
     if (have_neon(cpu_flags) && !high_bit_depth) {
-        /* c->put_h264_qpel_pixels_tab[0][ 0] = ff_put_h264_qpel16_mc00_neon; */
+        c->put_h264_qpel_pixels_tab[0][ 0] = ff_put_h264_qpel16_mc00_neon;
         c->put_h264_qpel_pixels_tab[0][ 1] = ff_put_h264_qpel16_mc10_neon;
         c->put_h264_qpel_pixels_tab[0][ 2] = ff_put_h264_qpel16_mc20_neon;
         c->put_h264_qpel_pixels_tab[0][ 3] = ff_put_h264_qpel16_mc30_neon;
@@ -118,7 +118,7 @@ av_cold void ff_h264qpel_init_aarch64(H264QpelContext *c, int bit_depth)
         c->put_h264_qpel_pixels_tab[0][14] = ff_put_h264_qpel16_mc23_neon;
         c->put_h264_qpel_pixels_tab[0][15] = ff_put_h264_qpel16_mc33_neon;
 
-        /* c->put_h264_qpel_pixels_tab[1][ 0] = ff_put_h264_qpel8_mc00_neon; */
+        c->put_h264_qpel_pixels_tab[1][ 0] = ff_put_h264_qpel8_mc00_neon;
         c->put_h264_qpel_pixels_tab[1][ 1] = ff_put_h264_qpel8_mc10_neon;
         c->put_h264_qpel_pixels_tab[1][ 2] = ff_put_h264_qpel8_mc20_neon;
         c->put_h264_qpel_pixels_tab[1][ 3] = ff_put_h264_qpel8_mc30_neon;
@@ -135,7 +135,7 @@ av_cold void ff_h264qpel_init_aarch64(H264QpelContext *c, int bit_depth)
         c->put_h264_qpel_pixels_tab[1][14] = ff_put_h264_qpel8_mc23_neon;
         c->put_h264_qpel_pixels_tab[1][15] = ff_put_h264_qpel8_mc33_neon;
 
-        /* c->avg_h264_qpel_pixels_tab[0][ 0] = ff_avg_h264_qpel16_mc00_neon; */
+        c->avg_h264_qpel_pixels_tab[0][ 0] = ff_avg_h264_qpel16_mc00_neon;
         c->avg_h264_qpel_pixels_tab[0][ 1] = ff_avg_h264_qpel16_mc10_neon;
         c->avg_h264_qpel_pixels_tab[0][ 2] = ff_avg_h264_qpel16_mc20_neon;
         c->avg_h264_qpel_pixels_tab[0][ 3] = ff_avg_h264_qpel16_mc30_neon;
@@ -152,7 +152,7 @@ av_cold void ff_h264qpel_init_aarch64(H264QpelContext *c, int bit_depth)
         c->avg_h264_qpel_pixels_tab[0][14] = ff_avg_h264_qpel16_mc23_neon;
         c->avg_h264_qpel_pixels_tab[0][15] = ff_avg_h264_qpel16_mc33_neon;
 
-        /* c->avg_h264_qpel_pixels_tab[1][ 0] = ff_avg_h264_qpel8_mc00_neon; */
+        c->avg_h264_qpel_pixels_tab[1][ 0] = ff_avg_h264_qpel8_mc00_neon;
         c->avg_h264_qpel_pixels_tab[1][ 1] = ff_avg_h264_qpel8_mc10_neon;
         c->avg_h264_qpel_pixels_tab[1][ 2] = ff_avg_h264_qpel8_mc20_neon;
         c->avg_h264_qpel_pixels_tab[1][ 3] = ff_avg_h264_qpel8_mc30_neon;

libavcodec/aarch64/hpeldsp_init_aarch64.c

+/*
+ * ARM NEON optimised DSP functions
+ * Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
+ *
+ * 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 <stddef.h>
+#include <stdint.h>
+
+#include "config.h"
+
+#include "libavutil/attributes.h"
+#include "libavutil/cpu.h"
+#include "libavutil/aarch64/cpu.h"
+#include "libavcodec/hpeldsp.h"
+
+void     ff_put_pixels16_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void  ff_put_pixels16_x2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void  ff_put_pixels16_y2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void ff_put_pixels16_xy2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void      ff_put_pixels8_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void   ff_put_pixels8_x2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void   ff_put_pixels8_y2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void  ff_put_pixels8_xy2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+
+void  ff_put_pixels16_x2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void  ff_put_pixels16_y2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void ff_put_pixels16_xy2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void   ff_put_pixels8_x2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void   ff_put_pixels8_y2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void  ff_put_pixels8_xy2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+
+void     ff_avg_pixels16_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void  ff_avg_pixels16_x2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void  ff_avg_pixels16_y2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void ff_avg_pixels16_xy2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void      ff_avg_pixels8_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void   ff_avg_pixels8_x2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void   ff_avg_pixels8_y2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+void  ff_avg_pixels8_xy2_neon(uint8_t *block, const uint8_t *pixels,
+                              ptrdiff_t line_size, int h);
+
+void  ff_avg_pixels16_x2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void  ff_avg_pixels16_y2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+void ff_avg_pixels16_xy2_no_rnd_neon(uint8_t *block, const uint8_t *pixels,
+                                     ptrdiff_t line_size, int h);
+
+av_cold void ff_hpeldsp_init_aarch64(HpelDSPContext *c, int flags)
+{
+    int cpu_flags = av_get_cpu_flags();
+
+    if (have_neon(cpu_flags)) {
+        c->put_pixels_tab[0][0] = ff_put_pixels16_neon;
+        c->put_pixels_tab[0][1] = ff_put_pixels16_x2_neon;
+        c->put_pixels_tab[0][2] = ff_put_pixels16_y2_neon;
+        c->put_pixels_tab[0][3] = ff_put_pixels16_xy2_neon;
+        c->put_pixels_tab[1][0] = ff_put_pixels8_neon;
+        c->put_pixels_tab[1][1] = ff_put_pixels8_x2_neon;
+        c->put_pixels_tab[1][2] = ff_put_pixels8_y2_neon;
+        c->put_pixels_tab[1][3] = ff_put_pixels8_xy2_neon;
+
+        c->put_no_rnd_pixels_tab[0][0] = ff_put_pixels16_neon;
+        c->put_no_rnd_pixels_tab[0][1] = ff_put_pixels16_x2_no_rnd_neon;
+        c->put_no_rnd_pixels_tab[0][2] = ff_put_pixels16_y2_no_rnd_neon;
+        c->put_no_rnd_pixels_tab[0][3] = ff_put_pixels16_xy2_no_rnd_neon;
+        c->put_no_rnd_pixels_tab[1][0] = ff_put_pixels8_neon;
+        c->put_no_rnd_pixels_tab[1][1] = ff_put_pixels8_x2_no_rnd_neon;
+        c->put_no_rnd_pixels_tab[1][2] = ff_put_pixels8_y2_no_rnd_neon;
+        c->put_no_rnd_pixels_tab[1][3] = ff_put_pixels8_xy2_no_rnd_neon;
+
+        c->avg_pixels_tab[0][0] = ff_avg_pixels16_neon;
+        c->avg_pixels_tab[0][1] = ff_avg_pixels16_x2_neon;
+        c->avg_pixels_tab[0][2] = ff_avg_pixels16_y2_neon;
+        c->avg_pixels_tab[0][3] = ff_avg_pixels16_xy2_neon;
+        c->avg_pixels_tab[1][0] = ff_avg_pixels8_neon;
+        c->avg_pixels_tab[1][1] = ff_avg_pixels8_x2_neon;
+        c->avg_pixels_tab[1][2] = ff_avg_pixels8_y2_neon;
+        c->avg_pixels_tab[1][3] = ff_avg_pixels8_xy2_neon;
+
+        c->avg_no_rnd_pixels_tab[0] = ff_avg_pixels16_neon;
+        c->avg_no_rnd_pixels_tab[1] = ff_avg_pixels16_x2_no_rnd_neon;
+        c->avg_no_rnd_pixels_tab[2] = ff_avg_pixels16_y2_no_rnd_neon;
+        c->avg_no_rnd_pixels_tab[3] = ff_avg_pixels16_xy2_no_rnd_neon;
+    }
+}

libavcodec/hpeldsp.c

@@ -54,6 +54,8 @@ av_cold void ff_hpeldsp_init(HpelDSPContext *c, int flags)
     hpel_funcs(avg, [3],  2);
     hpel_funcs(avg_no_rnd,, 16);
 
+    if (ARCH_AARCH64)
+        ff_hpeldsp_init_aarch64(c, flags);
     if (ARCH_ALPHA)
         ff_hpeldsp_init_alpha(c, flags);
     if (ARCH_ARM)

libavcodec/hpeldsp.h

@@ -94,6 +94,7 @@ typedef struct HpelDSPContext {
 
 void ff_hpeldsp_init(HpelDSPContext *c, int flags);
 
+void ff_hpeldsp_init_aarch64(HpelDSPContext *c, int flags);
 void ff_hpeldsp_init_alpha(HpelDSPContext *c, int flags);
 void ff_hpeldsp_init_arm(HpelDSPContext *c, int flags);
 void ff_hpeldsp_init_bfin(HpelDSPContext *c, int flags);