/*
* Copyright (c) 2004 Romain Dolbeau <romain@dolbeau.org>
*
- * This file is part of FFmpeg.
+ * This file is part of Libav.
*
- * FFmpeg is free software; you can redistribute it and/or
+ * 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.
*
- * FFmpeg is distributed in the hope that it will be useful,
+ * 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 FFmpeg; if not, write to the Free Software
+ * License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
+#include "libavutil/cpu.h"
#include "libavcodec/dsputil.h"
#include "libavcodec/h264data.h"
#include "libavcodec/h264dsp.h"
-#include "dsputil_ppc.h"
#include "dsputil_altivec.h"
#include "util_altivec.h"
#include "types_altivec.h"
#define OP_U8_ALTIVEC PUT_OP_U8_ALTIVEC
#define PREFIX_h264_chroma_mc8_altivec put_h264_chroma_mc8_altivec
-#define PREFIX_no_rnd_vc1_chroma_mc8_altivec put_no_rnd_vc1_chroma_mc8_altivec
#define PREFIX_h264_chroma_mc8_num altivec_put_h264_chroma_mc8_num
#define PREFIX_h264_qpel16_h_lowpass_altivec put_h264_qpel16_h_lowpass_altivec
#define PREFIX_h264_qpel16_h_lowpass_num altivec_put_h264_qpel16_h_lowpass_num
#include "h264_template_altivec.c"
#undef OP_U8_ALTIVEC
#undef PREFIX_h264_chroma_mc8_altivec
-#undef PREFIX_no_rnd_vc1_chroma_mc8_altivec
#undef PREFIX_h264_chroma_mc8_num
#undef PREFIX_h264_qpel16_h_lowpass_altivec
#undef PREFIX_h264_qpel16_h_lowpass_num
#define OP_U8_ALTIVEC AVG_OP_U8_ALTIVEC
#define PREFIX_h264_chroma_mc8_altivec avg_h264_chroma_mc8_altivec
-#define PREFIX_no_rnd_vc1_chroma_mc8_altivec avg_no_rnd_vc1_chroma_mc8_altivec
#define PREFIX_h264_chroma_mc8_num altivec_avg_h264_chroma_mc8_num
#define PREFIX_h264_qpel16_h_lowpass_altivec avg_h264_qpel16_h_lowpass_altivec
#define PREFIX_h264_qpel16_h_lowpass_num altivec_avg_h264_qpel16_h_lowpass_num
#include "h264_template_altivec.c"
#undef OP_U8_ALTIVEC
#undef PREFIX_h264_chroma_mc8_altivec
-#undef PREFIX_no_rnd_vc1_chroma_mc8_altivec
#undef PREFIX_h264_chroma_mc8_num
#undef PREFIX_h264_qpel16_h_lowpass_altivec
#undef PREFIX_h264_qpel16_h_lowpass_num
#define H264_MC(OPNAME, SIZE, CODETYPE) \
static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## CODETYPE (uint8_t *dst, uint8_t *src, int stride){\
- OPNAME ## pixels ## SIZE ## _ ## CODETYPE(dst, src, stride, SIZE);\
+ ff_ ## OPNAME ## pixels ## SIZE ## _ ## CODETYPE(dst, src, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){ \
h264_idct_dc_add_internal(dst, block, stride, 8);
}
-static void ff_h264_idct_add16_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+static void ff_h264_idct_add16_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
int i;
for(i=0; i<16; i++){
int nnz = nnzc[ scan8[i] ];
}
}
-static void ff_h264_idct_add16intra_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+static void ff_h264_idct_add16intra_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
int i;
for(i=0; i<16; i++){
if(nnzc[ scan8[i] ]) ff_h264_idct_add_altivec(dst + block_offset[i], block + i*16, stride);
}
}
-static void ff_h264_idct8_add4_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+static void ff_h264_idct8_add4_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
int i;
for(i=0; i<16; i+=4){
int nnz = nnzc[ scan8[i] ];
}
}
-static void ff_h264_idct_add8_altivec(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
- int i;
- for(i=16; i<16+8; i++){
- if(nnzc[ scan8[i] ])
- ff_h264_idct_add_altivec(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
- else if(block[i*16])
- h264_idct_dc_add_altivec(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
+static void ff_h264_idct_add8_altivec(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
+ int i, j;
+ for (j = 1; j < 3; j++) {
+ for(i = j * 16; i < j * 16 + 4; i++){
+ if(nnzc[ scan8[i] ])
+ ff_h264_idct_add_altivec(dest[j-1] + block_offset[i], block + i*16, stride);
+ else if(block[i*16])
+ h264_idct_dc_add_altivec(dest[j-1] + block_offset[i], block + i*16, stride);
+ }
}
}
*(dst_int+15*int_dst_stride) = *(src_int + 15);
}
-/** \brief performs a 6x16 transpose of data in src, and stores it to dst
- \todo FIXME: see if we can't spare some vec_lvsl() by them factorizing
+/** @brief performs a 6x16 transpose of data in src, and stores it to dst
+ @todo FIXME: see if we can't spare some vec_lvsl() by them factorizing
out of unaligned_load() */
#define readAndTranspose16x6(src, src_stride, r8, r9, r10, r11, r12, r13) {\
register vec_u8 r0 = unaligned_load(0, src); \
}
static av_always_inline
-void weight_h264_WxH_altivec(uint8_t *block, int stride, int log2_denom, int weight, int offset, int w, int h)
+void weight_h264_W_altivec(uint8_t *block, int stride, int height,
+ int log2_denom, int weight, int offset, int w)
{
int y, aligned;
vec_u8 vblock;
voffset = vec_splat(vtemp, 5);
aligned = !((unsigned long)block & 0xf);
- for (y=0; y<h; y++) {
+ for (y = 0; y < height; y++) {
vblock = vec_ld(0, block);
v0 = (vec_s16)vec_mergeh(zero_u8v, vblock);
}
static av_always_inline
-void biweight_h264_WxH_altivec(uint8_t *dst, uint8_t *src, int stride, int log2_denom,
- int weightd, int weights, int offset, int w, int h)
+void biweight_h264_W_altivec(uint8_t *dst, uint8_t *src, int stride, int height,
+ int log2_denom, int weightd, int weights, int offset, int w)
{
int y, dst_aligned, src_aligned;
vec_u8 vsrc, vdst;
dst_aligned = !((unsigned long)dst & 0xf);
src_aligned = !((unsigned long)src & 0xf);
- for (y=0; y<h; y++) {
+ for (y = 0; y < height; y++) {
vdst = vec_ld(0, dst);
vsrc = vec_ld(0, src);
}
}
-#define H264_WEIGHT(W,H) \
-static void ff_weight_h264_pixels ## W ## x ## H ## _altivec(uint8_t *block, int stride, int log2_denom, int weight, int offset){ \
- weight_h264_WxH_altivec(block, stride, log2_denom, weight, offset, W, H); \
+#define H264_WEIGHT(W) \
+static void ff_weight_h264_pixels ## W ## _altivec(uint8_t *block, int stride, int height, \
+ int log2_denom, int weight, int offset){ \
+ weight_h264_W_altivec(block, stride, height, log2_denom, weight, offset, W); \
}\
-static void ff_biweight_h264_pixels ## W ## x ## H ## _altivec(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset){ \
- biweight_h264_WxH_altivec(dst, src, stride, log2_denom, weightd, weights, offset, W, H); \
+static void ff_biweight_h264_pixels ## W ## _altivec(uint8_t *dst, uint8_t *src, int stride, int height, \
+ int log2_denom, int weightd, int weights, int offset){ \
+ biweight_h264_W_altivec(dst, src, stride, height, log2_denom, weightd, weights, offset, W); \
}
-H264_WEIGHT(16,16)
-H264_WEIGHT(16, 8)
-H264_WEIGHT( 8,16)
-H264_WEIGHT( 8, 8)
-H264_WEIGHT( 8, 4)
+H264_WEIGHT(16)
+H264_WEIGHT( 8)
-void dsputil_h264_init_ppc(DSPContext* c, AVCodecContext *avctx) {
+void ff_dsputil_h264_init_ppc(DSPContext* c, AVCodecContext *avctx) {
+ const int high_bit_depth = avctx->bits_per_raw_sample > 8;
- if (has_altivec()) {
+ if (av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC) {
+ if (!high_bit_depth) {
c->put_h264_chroma_pixels_tab[0] = put_h264_chroma_mc8_altivec;
c->avg_h264_chroma_pixels_tab[0] = avg_h264_chroma_mc8_altivec;
- c->put_no_rnd_vc1_chroma_pixels_tab[0] = put_no_rnd_vc1_chroma_mc8_altivec;
- c->avg_no_rnd_vc1_chroma_pixels_tab[0] = avg_no_rnd_vc1_chroma_mc8_altivec;
#define dspfunc(PFX, IDX, NUM) \
c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_altivec; \
dspfunc(avg_h264_qpel, 0, 16);
#undef dspfunc
}
+ }
}
-void ff_h264dsp_init_ppc(H264DSPContext *c)
+void ff_h264dsp_init_ppc(H264DSPContext *c, const int bit_depth, const int chroma_format_idc)
{
- if (has_altivec()) {
+ if (av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC) {
+ if (bit_depth == 8) {
c->h264_idct_add = ff_h264_idct_add_altivec;
- c->h264_idct_add8 = ff_h264_idct_add8_altivec;
+ if (chroma_format_idc == 1)
+ c->h264_idct_add8 = ff_h264_idct_add8_altivec;
c->h264_idct_add16 = ff_h264_idct_add16_altivec;
c->h264_idct_add16intra = ff_h264_idct_add16intra_altivec;
c->h264_idct_dc_add= h264_idct_dc_add_altivec;
c->h264_v_loop_filter_luma= h264_v_loop_filter_luma_altivec;
c->h264_h_loop_filter_luma= h264_h_loop_filter_luma_altivec;
- c->weight_h264_pixels_tab[0] = ff_weight_h264_pixels16x16_altivec;
- c->weight_h264_pixels_tab[1] = ff_weight_h264_pixels16x8_altivec;
- c->weight_h264_pixels_tab[2] = ff_weight_h264_pixels8x16_altivec;
- c->weight_h264_pixels_tab[3] = ff_weight_h264_pixels8x8_altivec;
- c->weight_h264_pixels_tab[4] = ff_weight_h264_pixels8x4_altivec;
- c->biweight_h264_pixels_tab[0] = ff_biweight_h264_pixels16x16_altivec;
- c->biweight_h264_pixels_tab[1] = ff_biweight_h264_pixels16x8_altivec;
- c->biweight_h264_pixels_tab[2] = ff_biweight_h264_pixels8x16_altivec;
- c->biweight_h264_pixels_tab[3] = ff_biweight_h264_pixels8x8_altivec;
- c->biweight_h264_pixels_tab[4] = ff_biweight_h264_pixels8x4_altivec;
+ c->weight_h264_pixels_tab[0] = ff_weight_h264_pixels16_altivec;
+ c->weight_h264_pixels_tab[1] = ff_weight_h264_pixels8_altivec;
+ c->biweight_h264_pixels_tab[0] = ff_biweight_h264_pixels16_altivec;
+ c->biweight_h264_pixels_tab[1] = ff_biweight_h264_pixels8_altivec;
+ }
}
}