FFmpeg: libavcodec/arm/rv40dsp_init

00001 /*
00002  * Copyright (c) 2011 Janne Grunau <janne-libav@jannau.net>
00003  *
00004  * This file is part of FFmpeg.
00005  *
00006  * FFmpeg is free software; you can redistribute it and/or
00007  * modify it under the terms of the GNU Lesser General Public
00008  * License as published by the Free Software Foundation; either
00009  * version 2.1 of the License, or (at your option) any later version.
00010  *
00011  * FFmpeg is distributed in the hope that it will be useful,
00012  * but WITHOUT ANY WARRANTY; without even the implied warranty of
00013  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
00014  * Lesser General Public License for more details.
00015  *
00016  * You should have received a copy of the GNU Lesser General Public
00017  * License along with FFmpeg; if not, write to the Free Software
00018  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
00019  */
00020 
00021 #include <stdint.h>
00022 
00023 #include "libavcodec/avcodec.h"
00024 #include "libavcodec/rv34dsp.h"
00025 #include "libavutil/arm/cpu.h"
00026 
00027 #define DECL_QPEL3(type, w, pos) \
00028     void ff_##type##_rv40_qpel##w##_mc##pos##_neon(uint8_t *dst, uint8_t *src,\
00029                                                    int stride)
00030 #define DECL_QPEL2(w, pos)                      \
00031     DECL_QPEL3(put, w, pos);                    \
00032     DECL_QPEL3(avg, w, pos)
00033 
00034 #define DECL_QPEL_XY(x, y)                      \
00035     DECL_QPEL2(16, x ## y);                     \
00036     DECL_QPEL2(8,  x ## y)
00037 
00038 #define DECL_QPEL_Y(y)                          \
00039     DECL_QPEL_XY(0, y);                         \
00040     DECL_QPEL_XY(1, y);                         \
00041     DECL_QPEL_XY(2, y);                         \
00042     DECL_QPEL_XY(3, y);                         \
00043 
00044 DECL_QPEL_Y(0);
00045 DECL_QPEL_Y(1);
00046 DECL_QPEL_Y(2);
00047 DECL_QPEL_Y(3);
00048 
00049 void ff_put_rv40_chroma_mc8_neon(uint8_t *, uint8_t *, int, int, int, int);
00050 void ff_put_rv40_chroma_mc4_neon(uint8_t *, uint8_t *, int, int, int, int);
00051 
00052 void ff_avg_rv40_chroma_mc8_neon(uint8_t *, uint8_t *, int, int, int, int);
00053 void ff_avg_rv40_chroma_mc4_neon(uint8_t *, uint8_t *, int, int, int, int);
00054 
00055 void ff_rv40_weight_func_16_neon(uint8_t *, uint8_t *, uint8_t *, int, int, ptrdiff_t);
00056 void ff_rv40_weight_func_8_neon(uint8_t *, uint8_t *, uint8_t *, int, int, ptrdiff_t);
00057 
00058 int ff_rv40_h_loop_filter_strength_neon(uint8_t *src, ptrdiff_t stride,
00059                                         int beta, int beta2, int edge,
00060                                         int *p1, int *q1);
00061 int ff_rv40_v_loop_filter_strength_neon(uint8_t *src, ptrdiff_t stride,
00062                                         int beta, int beta2, int edge,
00063                                         int *p1, int *q1);
00064 
00065 void ff_rv40_h_weak_loop_filter_neon(uint8_t *src, ptrdiff_t stride, int filter_p1,
00066                                      int filter_q1, int alpha, int beta,
00067                                      int lim_p0q0, int lim_q1, int lim_p1);
00068 void ff_rv40_v_weak_loop_filter_neon(uint8_t *src, ptrdiff_t stride, int filter_p1,
00069                                      int filter_q1, int alpha, int beta,
00070                                      int lim_p0q0, int lim_q1, int lim_p1);
00071 
00072 static void ff_rv40dsp_init_neon(RV34DSPContext *c)
00073 {
00074     c->put_pixels_tab[0][ 1] = ff_put_rv40_qpel16_mc10_neon;
00075     c->put_pixels_tab[0][ 3] = ff_put_rv40_qpel16_mc30_neon;
00076     c->put_pixels_tab[0][ 4] = ff_put_rv40_qpel16_mc01_neon;
00077     c->put_pixels_tab[0][ 5] = ff_put_rv40_qpel16_mc11_neon;
00078     c->put_pixels_tab[0][ 6] = ff_put_rv40_qpel16_mc21_neon;
00079     c->put_pixels_tab[0][ 7] = ff_put_rv40_qpel16_mc31_neon;
00080     c->put_pixels_tab[0][ 9] = ff_put_rv40_qpel16_mc12_neon;
00081     c->put_pixels_tab[0][10] = ff_put_rv40_qpel16_mc22_neon;
00082     c->put_pixels_tab[0][11] = ff_put_rv40_qpel16_mc32_neon;
00083     c->put_pixels_tab[0][12] = ff_put_rv40_qpel16_mc03_neon;
00084     c->put_pixels_tab[0][13] = ff_put_rv40_qpel16_mc13_neon;
00085     c->put_pixels_tab[0][14] = ff_put_rv40_qpel16_mc23_neon;
00086     c->put_pixels_tab[0][15] = ff_put_rv40_qpel16_mc33_neon;
00087     c->avg_pixels_tab[0][ 1] = ff_avg_rv40_qpel16_mc10_neon;
00088     c->avg_pixels_tab[0][ 3] = ff_avg_rv40_qpel16_mc30_neon;
00089     c->avg_pixels_tab[0][ 4] = ff_avg_rv40_qpel16_mc01_neon;
00090     c->avg_pixels_tab[0][ 5] = ff_avg_rv40_qpel16_mc11_neon;
00091     c->avg_pixels_tab[0][ 6] = ff_avg_rv40_qpel16_mc21_neon;
00092     c->avg_pixels_tab[0][ 7] = ff_avg_rv40_qpel16_mc31_neon;
00093     c->avg_pixels_tab[0][ 9] = ff_avg_rv40_qpel16_mc12_neon;
00094     c->avg_pixels_tab[0][10] = ff_avg_rv40_qpel16_mc22_neon;
00095     c->avg_pixels_tab[0][11] = ff_avg_rv40_qpel16_mc32_neon;
00096     c->avg_pixels_tab[0][12] = ff_avg_rv40_qpel16_mc03_neon;
00097     c->avg_pixels_tab[0][13] = ff_avg_rv40_qpel16_mc13_neon;
00098     c->avg_pixels_tab[0][14] = ff_avg_rv40_qpel16_mc23_neon;
00099     c->avg_pixels_tab[0][15] = ff_avg_rv40_qpel16_mc33_neon;
00100     c->put_pixels_tab[1][ 1] = ff_put_rv40_qpel8_mc10_neon;
00101     c->put_pixels_tab[1][ 3] = ff_put_rv40_qpel8_mc30_neon;
00102     c->put_pixels_tab[1][ 4] = ff_put_rv40_qpel8_mc01_neon;
00103     c->put_pixels_tab[1][ 5] = ff_put_rv40_qpel8_mc11_neon;
00104     c->put_pixels_tab[1][ 6] = ff_put_rv40_qpel8_mc21_neon;
00105     c->put_pixels_tab[1][ 7] = ff_put_rv40_qpel8_mc31_neon;
00106     c->put_pixels_tab[1][ 9] = ff_put_rv40_qpel8_mc12_neon;
00107     c->put_pixels_tab[1][10] = ff_put_rv40_qpel8_mc22_neon;
00108     c->put_pixels_tab[1][11] = ff_put_rv40_qpel8_mc32_neon;
00109     c->put_pixels_tab[1][12] = ff_put_rv40_qpel8_mc03_neon;
00110     c->put_pixels_tab[1][13] = ff_put_rv40_qpel8_mc13_neon;
00111     c->put_pixels_tab[1][14] = ff_put_rv40_qpel8_mc23_neon;
00112     c->put_pixels_tab[1][15] = ff_put_rv40_qpel8_mc33_neon;
00113     c->avg_pixels_tab[1][ 1] = ff_avg_rv40_qpel8_mc10_neon;
00114     c->avg_pixels_tab[1][ 3] = ff_avg_rv40_qpel8_mc30_neon;
00115     c->avg_pixels_tab[1][ 4] = ff_avg_rv40_qpel8_mc01_neon;
00116     c->avg_pixels_tab[1][ 5] = ff_avg_rv40_qpel8_mc11_neon;
00117     c->avg_pixels_tab[1][ 6] = ff_avg_rv40_qpel8_mc21_neon;
00118     c->avg_pixels_tab[1][ 7] = ff_avg_rv40_qpel8_mc31_neon;
00119     c->avg_pixels_tab[1][ 9] = ff_avg_rv40_qpel8_mc12_neon;
00120     c->avg_pixels_tab[1][10] = ff_avg_rv40_qpel8_mc22_neon;
00121     c->avg_pixels_tab[1][11] = ff_avg_rv40_qpel8_mc32_neon;
00122     c->avg_pixels_tab[1][12] = ff_avg_rv40_qpel8_mc03_neon;
00123     c->avg_pixels_tab[1][13] = ff_avg_rv40_qpel8_mc13_neon;
00124     c->avg_pixels_tab[1][14] = ff_avg_rv40_qpel8_mc23_neon;
00125     c->avg_pixels_tab[1][15] = ff_avg_rv40_qpel8_mc33_neon;
00126 
00127     c->put_chroma_pixels_tab[0] = ff_put_rv40_chroma_mc8_neon;
00128     c->put_chroma_pixels_tab[1] = ff_put_rv40_chroma_mc4_neon;
00129     c->avg_chroma_pixels_tab[0] = ff_avg_rv40_chroma_mc8_neon;
00130     c->avg_chroma_pixels_tab[1] = ff_avg_rv40_chroma_mc4_neon;
00131 
00132     c->rv40_weight_pixels_tab[0][0] = ff_rv40_weight_func_16_neon;
00133     c->rv40_weight_pixels_tab[0][1] = ff_rv40_weight_func_8_neon;
00134 
00135     c->rv40_loop_filter_strength[0] = ff_rv40_h_loop_filter_strength_neon;
00136     c->rv40_loop_filter_strength[1] = ff_rv40_v_loop_filter_strength_neon;
00137     c->rv40_weak_loop_filter[0]     = ff_rv40_h_weak_loop_filter_neon;
00138     c->rv40_weak_loop_filter[1]     = ff_rv40_v_weak_loop_filter_neon;
00139 }
00140 
00141 void ff_rv40dsp_init_arm(RV34DSPContext *c, DSPContext* dsp)
00142 {
00143     int cpu_flags = av_get_cpu_flags();
00144 
00145     if (have_neon(cpu_flags))
00146         ff_rv40dsp_init_neon(c);
00147 }
libavcodec/arm/rv40dsp_init_arm.c
