/*****************************************************************************
- * mc-c.c: h264 encoder library (Motion Compensation)
+ * mc-c.c: arm motion compensation
*****************************************************************************
- * Copyright (C) 2009 x264 project
+ * Copyright (C) 2009-2014 x264 project
*
* Authors: David Conrad <lessen42@gmail.com>
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
+ *
+ * This program is also available under a commercial proprietary license.
+ * For more information, contact us at licensing@x264.com.
*****************************************************************************/
#include "common/common.h"
#include "mc.h"
-void x264_prefetch_ref_arm( uint8_t *, int, int );
-void x264_prefetch_fenc_arm( uint8_t *, int, uint8_t *, int, int );
+void x264_prefetch_ref_arm( uint8_t *, intptr_t, int );
+void x264_prefetch_fenc_arm( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+
+void *x264_memcpy_aligned_neon( void *dst, const void *src, size_t n );
+void x264_memzero_aligned_neon( void *dst, size_t n );
+
+void x264_pixel_avg_16x16_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_16x8_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_8x16_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_8x8_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_8x4_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_4x16_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_4x8_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_4x4_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_pixel_avg_4x2_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, intptr_t, int );
-void *x264_memcpy_aligned_neon( void * dst, const void * src, size_t n );
-void x264_memzero_aligned_neon( void *dst, int n );
+void x264_pixel_avg2_w4_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
+void x264_pixel_avg2_w8_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
+void x264_pixel_avg2_w16_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
+void x264_pixel_avg2_w20_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int );
-void x264_pixel_avg_16x16_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_16x8_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_8x16_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_8x8_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_8x4_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_4x8_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_4x4_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
-void x264_pixel_avg_4x2_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int, int );
+void x264_plane_copy_deinterleave_neon( pixel *dstu, intptr_t i_dstu,
+ pixel *dstv, intptr_t i_dstv,
+ pixel *src, intptr_t i_src, int w, int h );
+void x264_plane_copy_deinterleave_rgb_neon( pixel *dsta, intptr_t i_dsta,
+ pixel *dstb, intptr_t i_dstb,
+ pixel *dstc, intptr_t i_dstc,
+ pixel *src, intptr_t i_src, int pw, int w, int h );
+void x264_plane_copy_interleave_neon( pixel *dst, intptr_t i_dst,
+ pixel *srcu, intptr_t i_srcu,
+ pixel *srcv, intptr_t i_srcv, int w, int h );
-void x264_pixel_avg2_w4_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int );
-void x264_pixel_avg2_w8_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int );
-void x264_pixel_avg2_w16_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int );
-void x264_pixel_avg2_w20_neon( uint8_t *, int, uint8_t *, int, uint8_t *, int );
+void x264_store_interleave_chroma_neon( pixel *dst, intptr_t i_dst, pixel *srcu, pixel *srcv, int height );
+void x264_load_deinterleave_chroma_fdec_neon( pixel *dst, pixel *src, intptr_t i_src, int height );
+void x264_load_deinterleave_chroma_fenc_neon( pixel *dst, pixel *src, intptr_t i_src, int height );
#define MC_WEIGHT(func)\
-void x264_mc_weight_w20##func##_neon( uint8_t *, int, uint8_t *, int, const x264_weight_t *, int );\
-void x264_mc_weight_w16##func##_neon( uint8_t *, int, uint8_t *, int, const x264_weight_t *, int );\
-void x264_mc_weight_w8##func##_neon( uint8_t *, int, uint8_t *, int, const x264_weight_t *, int );\
-void x264_mc_weight_w4##func##_neon( uint8_t *, int, uint8_t *, int, const x264_weight_t *, int );\
+void x264_mc_weight_w20##func##_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, const x264_weight_t *, int );\
+void x264_mc_weight_w16##func##_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, const x264_weight_t *, int );\
+void x264_mc_weight_w8##func##_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, const x264_weight_t *, int );\
+void x264_mc_weight_w4##func##_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, const x264_weight_t *, int );\
\
-static void (* const x264_mc##func##_wtab_neon[6])( uint8_t *, int, uint8_t *, int, const x264_weight_t *, int ) =\
+static void (* const x264_mc##func##_wtab_neon[6])( uint8_t *, intptr_t, uint8_t *, intptr_t, const x264_weight_t *, int ) =\
{\
x264_mc_weight_w4##func##_neon,\
x264_mc_weight_w4##func##_neon,\
MC_WEIGHT(_offsetadd)
MC_WEIGHT(_offsetsub)
-void x264_mc_copy_w4_neon( uint8_t *, int, uint8_t *, int, int );
-void x264_mc_copy_w8_neon( uint8_t *, int, uint8_t *, int, int );
-void x264_mc_copy_w16_neon( uint8_t *, int, uint8_t *, int, int );
-void x264_mc_copy_w16_aligned_neon( uint8_t *, int, uint8_t *, int, int );
+void x264_mc_copy_w4_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_mc_copy_w8_neon ( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_mc_copy_w16_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
+void x264_mc_copy_w16_aligned_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, int );
-void x264_mc_chroma_neon( uint8_t *, int, uint8_t *, int, int, int, int, int );
-void x264_frame_init_lowres_core_neon( uint8_t *, uint8_t *, uint8_t *, uint8_t *, uint8_t *, int, int, int, int);
+void x264_mc_chroma_neon( uint8_t *, intptr_t, uint8_t *, intptr_t, int, int, int, int );
+void x264_frame_init_lowres_core_neon( uint8_t *, uint8_t *, uint8_t *, uint8_t *, uint8_t *, intptr_t, intptr_t, int, int );
-void x264_hpel_filter_v_neon( uint8_t *, uint8_t *, int16_t *, int, int );
+void x264_hpel_filter_v_neon( uint8_t *, uint8_t *, int16_t *, intptr_t, int );
void x264_hpel_filter_c_neon( uint8_t *, int16_t *, int );
void x264_hpel_filter_h_neon( uint8_t *, uint8_t *, int );
-#if !X264_HIGH_BIT_DEPTH
+#if !HIGH_BIT_DEPTH
static void x264_weight_cache_neon( x264_t *h, x264_weight_t *w )
{
if( w->i_scale == 1<<w->i_denom )
w->weightfn = x264_mc_wtab_neon;
}
-static void (* const x264_pixel_avg_wtab_neon[6])( uint8_t *, int, uint8_t *, int, uint8_t *, int ) =
+static void (* const x264_pixel_avg_wtab_neon[6])( uint8_t *, intptr_t, uint8_t *, intptr_t, uint8_t *, int ) =
{
NULL,
x264_pixel_avg2_w4_neon,
x264_pixel_avg2_w20_neon,
};
-static void (* const x264_mc_copy_wtab_neon[5])( uint8_t *, int, uint8_t *, int, int ) =
+static void (* const x264_mc_copy_wtab_neon[5])( uint8_t *, intptr_t, uint8_t *, intptr_t, int ) =
{
NULL,
x264_mc_copy_w4_neon,
static const uint8_t hpel_ref0[16] = {0,1,1,1,0,1,1,1,2,3,3,3,0,1,1,1};
static const uint8_t hpel_ref1[16] = {0,0,0,0,2,2,3,2,2,2,3,2,2,2,3,2};
-static void mc_luma_neon( uint8_t *dst, int i_dst_stride,
- uint8_t *src[4], int i_src_stride,
+static void mc_luma_neon( uint8_t *dst, intptr_t i_dst_stride,
+ uint8_t *src[4], intptr_t i_src_stride,
int mvx, int mvy,
int i_width, int i_height, const x264_weight_t *weight )
{
int qpel_idx = ((mvy&3)<<2) + (mvx&3);
- int offset = (mvy>>2)*i_src_stride + (mvx>>2);
+ intptr_t offset = (mvy>>2)*i_src_stride + (mvx>>2);
uint8_t *src1 = src[hpel_ref0[qpel_idx]] + offset;
if ( (mvy&3) == 3 ) // explict if() to force conditional add
src1 += i_src_stride;
x264_mc_copy_wtab_neon[i_width>>2]( dst, i_dst_stride, src1, i_src_stride, i_height );
}
-static uint8_t *get_ref_neon( uint8_t *dst, int *i_dst_stride,
- uint8_t *src[4], int i_src_stride,
+static uint8_t *get_ref_neon( uint8_t *dst, intptr_t *i_dst_stride,
+ uint8_t *src[4], intptr_t i_src_stride,
int mvx, int mvy,
int i_width, int i_height, const x264_weight_t *weight )
{
int qpel_idx = ((mvy&3)<<2) + (mvx&3);
- int offset = (mvy>>2)*i_src_stride + (mvx>>2);
+ intptr_t offset = (mvy>>2)*i_src_stride + (mvx>>2);
uint8_t *src1 = src[hpel_ref0[qpel_idx]] + offset;
if ( (mvy&3) == 3 ) // explict if() to force conditional add
src1 += i_src_stride;
}
static void hpel_filter_neon( uint8_t *dsth, uint8_t *dstv, uint8_t *dstc, uint8_t *src,
- int stride, int width, int height, int16_t *buf )
+ intptr_t stride, int width, int height, int16_t *buf )
{
- int realign = (intptr_t)src & 15;
+ intptr_t realign = (intptr_t)src & 15;
src -= realign;
dstv -= realign;
dstc -= realign;
src += stride;
}
}
-#endif // !X264_HIGH_BIT_DEPTH
+#endif // !HIGH_BIT_DEPTH
void x264_mc_init_arm( int cpu, x264_mc_functions_t *pf )
{
if( !(cpu&X264_CPU_ARMV6) )
return;
-#if !X264_HIGH_BIT_DEPTH
- pf->prefetch_fenc = x264_prefetch_fenc_arm;
+#if !HIGH_BIT_DEPTH
+ pf->prefetch_fenc_420 = x264_prefetch_fenc_arm;
+ pf->prefetch_fenc_422 = x264_prefetch_fenc_arm; /* FIXME */
pf->prefetch_ref = x264_prefetch_ref_arm;
-#endif // !X264_HIGH_BIT_DEPTH
+#endif // !HIGH_BIT_DEPTH
if( !(cpu&X264_CPU_NEON) )
return;
-#if !X264_HIGH_BIT_DEPTH
+#if !HIGH_BIT_DEPTH
pf->copy_16x16_unaligned = x264_mc_copy_w16_neon;
pf->copy[PIXEL_16x16] = x264_mc_copy_w16_aligned_neon;
pf->copy[PIXEL_8x8] = x264_mc_copy_w8_neon;
pf->copy[PIXEL_4x4] = x264_mc_copy_w4_neon;
+ pf->plane_copy_deinterleave = x264_plane_copy_deinterleave_neon;
+ pf->plane_copy_deinterleave_rgb = x264_plane_copy_deinterleave_rgb_neon;
+ pf->plane_copy_interleave = x264_plane_copy_interleave_neon;
+
+ pf->store_interleave_chroma = x264_store_interleave_chroma_neon;
+ pf->load_deinterleave_chroma_fdec = x264_load_deinterleave_chroma_fdec_neon;
+ pf->load_deinterleave_chroma_fenc = x264_load_deinterleave_chroma_fenc_neon;
+
pf->avg[PIXEL_16x16] = x264_pixel_avg_16x16_neon;
pf->avg[PIXEL_16x8] = x264_pixel_avg_16x8_neon;
pf->avg[PIXEL_8x16] = x264_pixel_avg_8x16_neon;
pf->avg[PIXEL_8x8] = x264_pixel_avg_8x8_neon;
pf->avg[PIXEL_8x4] = x264_pixel_avg_8x4_neon;
+ pf->avg[PIXEL_4x16] = x264_pixel_avg_4x16_neon;
pf->avg[PIXEL_4x8] = x264_pixel_avg_4x8_neon;
pf->avg[PIXEL_4x4] = x264_pixel_avg_4x4_neon;
pf->avg[PIXEL_4x2] = x264_pixel_avg_4x2_neon;
pf->get_ref = get_ref_neon;
pf->hpel_filter = hpel_filter_neon;
pf->frame_init_lowres_core = x264_frame_init_lowres_core_neon;
-#endif // !X264_HIGH_BIT_DEPTH
+#endif // !HIGH_BIT_DEPTH
// Apple's gcc stupidly cannot align stack variables, and ALIGNED_ARRAY can't work on structs
#ifndef SYS_MACOSX