virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-01-29 22:00:58 +02:00
Code Issues Releases Activity

avcodec/mpegvideo: Inline is_encoder in mpv_reconstruct_mb_internal()

Browse Source 
Up until now, we inlined lowres_flag as well as is_mpeg12
independently (unless CONFIG_SMALL was true); this commit
changes this to instead inline mpv_reconstruct_mb_internal()
(at most) four times, namely once for encoders, once for decoders
using lowres and once for non-lowres mpeg-1/2 decoders and once
for non-lowres non-mpeg-1/2 decoders (mpeg-1/2 is not inlined
in case of CONFIG_SMALL). This is neutral performance-wise,
but proved beneficial size-wise: It saved 1776B of .text
for GCC 11 or 1344B for Clang 14 (both -O3 x64).

Notice that inlining is_mpeg12 for is_encoder would not really
be beneficial, as the encoder codepath does mostly not depend
on is_mpeg12 at all.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This commit is contained in:
Andreas Rheinhardt 2022-10-14 04:11:34 +02:00
parent 409c4723ec
commit 9ca312d8ab
1 changed files with 18 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
libavcodec/mpegvideo.c
View File

@ -1310,9 +1310,8 @@ void ff_clean_intra_table_entries(MpegEncContext *s)
*/ */
static av_always_inline static av_always_inline
void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64], void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64],
int lowres_flag, int is_mpeg12) int lowres_flag, int is_mpeg12, int is_encoder)
{ {
#define IS_ENCODER(s) (CONFIG_MPEGVIDEOENC && !lowres_flag && (s)->encoding)
#define IS_MPEG12(s) (is_mpeg12 == MAY_BE_MPEG12 ? ((s)->out_format == FMT_MPEG1) : is_mpeg12) #define IS_MPEG12(s) (is_mpeg12 == MAY_BE_MPEG12 ? ((s)->out_format == FMT_MPEG1) : is_mpeg12)
const int mb_xy = s->mb_y * s->mb_stride + s->mb_x; const int mb_xy = s->mb_y * s->mb_stride + s->mb_x;
@ -1331,7 +1330,7 @@ void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64],
} else if (is_mpeg12 != DEFINITELY_MPEG12 && (s->h263_pred || s->h263_aic)) } else if (is_mpeg12 != DEFINITELY_MPEG12 && (s->h263_pred || s->h263_aic))
s->mbintra_table[mb_xy]=1; s->mbintra_table[mb_xy]=1;
if (!IS_ENCODER(s) || (s->avctx->flags & AV_CODEC_FLAG_PSNR) || s->frame_skip_threshold || s->frame_skip_factor || if (!is_encoder || (s->avctx->flags & AV_CODEC_FLAG_PSNR) || s->frame_skip_threshold || s->frame_skip_factor ||
!((s->intra_only || s->pict_type == AV_PICTURE_TYPE_B) && !((s->intra_only || s->pict_type == AV_PICTURE_TYPE_B) &&
s->avctx->mb_decision != FF_MB_DECISION_RD)) { // FIXME precalc s->avctx->mb_decision != FF_MB_DECISION_RD)) { // FIXME precalc
uint8_t *dest_y, *dest_cb, *dest_cr; uint8_t *dest_y, *dest_cb, *dest_cr;
@ -1340,12 +1339,12 @@ void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64],
qpel_mc_func (*op_qpix)[16]; qpel_mc_func (*op_qpix)[16];
const int linesize = s->current_picture.f->linesize[0]; //not s->linesize as this would be wrong for field pics const int linesize = s->current_picture.f->linesize[0]; //not s->linesize as this would be wrong for field pics
const int uvlinesize = s->current_picture.f->linesize[1]; const int uvlinesize = s->current_picture.f->linesize[1];
const int readable = s->pict_type != AV_PICTURE_TYPE_B || IS_ENCODER(s) || lowres_flag; const int readable = s->pict_type != AV_PICTURE_TYPE_B || is_encoder || lowres_flag;
const int block_size= lowres_flag ? 8>>s->avctx->lowres : 8; const int block_size= lowres_flag ? 8>>s->avctx->lowres : 8;
/* avoid copy if macroblock skipped in last frame too */ /* avoid copy if macroblock skipped in last frame too */
/* skip only during decoding as we might trash the buffers during encoding a bit */ /* skip only during decoding as we might trash the buffers during encoding a bit */
if (!IS_ENCODER(s)) { if (!is_encoder) {
uint8_t *mbskip_ptr = &s->mbskip_table[mb_xy]; uint8_t *mbskip_ptr = &s->mbskip_table[mb_xy];
if (s->mb_skipped) { if (s->mb_skipped) {
@ -1375,7 +1374,7 @@ void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64],
if (!s->mb_intra) { if (!s->mb_intra) {
/* motion handling */ /* motion handling */
/* decoding or more than one mb_type (MC was already done otherwise) */ /* decoding or more than one mb_type (MC was already done otherwise) */
if (!IS_ENCODER(s)) { if (!is_encoder) {
if(HAVE_THREADS && s->avctx->active_thread_type&FF_THREAD_FRAME) { if(HAVE_THREADS && s->avctx->active_thread_type&FF_THREAD_FRAME) {
if (s->mv_dir & MV_DIR_FORWARD) { if (s->mv_dir & MV_DIR_FORWARD) {
@ -1427,7 +1426,7 @@ void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64],
} }
/* add dct residue */ /* add dct residue */
if (IS_ENCODER(s) || !(IS_MPEG12(s) || s->msmpeg4_version if (is_encoder || !(IS_MPEG12(s) || s->msmpeg4_version
|| (s->codec_id==AV_CODEC_ID_MPEG4 && !s->mpeg_quant))){ || (s->codec_id==AV_CODEC_ID_MPEG4 && !s->mpeg_quant))){
add_dequant_dct(s, block[0], 0, dest_y , dct_linesize, s->qscale); add_dequant_dct(s, block[0], 0, dest_y , dct_linesize, s->qscale);
add_dequant_dct(s, block[1], 1, dest_y + block_size, dct_linesize, s->qscale); add_dequant_dct(s, block[1], 1, dest_y + block_size, dct_linesize, s->qscale);
@ -1480,14 +1479,14 @@ void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64],
} else { } else {
/* Only MPEG-4 Simple Studio Profile is supported in > 8-bit mode. /* Only MPEG-4 Simple Studio Profile is supported in > 8-bit mode.
TODO: Integrate 10-bit properly into mpegvideo.c so that ER works properly */ TODO: Integrate 10-bit properly into mpegvideo.c so that ER works properly */
if (is_mpeg12 != DEFINITELY_MPEG12 && CONFIG_MPEG4_DECODER && if (!is_encoder && is_mpeg12 != DEFINITELY_MPEG12 && CONFIG_MPEG4_DECODER &&
/* s->codec_id == AV_CODEC_ID_MPEG4 && */ /* s->codec_id == AV_CODEC_ID_MPEG4 && */
s->avctx->bits_per_raw_sample > 8) { s->avctx->bits_per_raw_sample > 8) {
ff_mpeg4_decode_studio(s, dest_y, dest_cb, dest_cr, block_size, ff_mpeg4_decode_studio(s, dest_y, dest_cb, dest_cr, block_size,
uvlinesize, dct_linesize, dct_offset); uvlinesize, dct_linesize, dct_offset);
} }
/* dct only in intra block */ /* dct only in intra block */
else if (IS_ENCODER(s) || !IS_MPEG12(s)) { else if (is_encoder || !IS_MPEG12(s)) {
put_dct(s, block[0], 0, dest_y , dct_linesize, s->qscale); put_dct(s, block[0], 0, dest_y , dct_linesize, s->qscale);
put_dct(s, block[1], 1, dest_y + block_size, dct_linesize, s->qscale); put_dct(s, block[1], 1, dest_y + block_size, dct_linesize, s->qscale);
put_dct(s, block[2], 2, dest_y + dct_offset , dct_linesize, s->qscale); put_dct(s, block[2], 2, dest_y + dct_offset , dct_linesize, s->qscale);
@ -1560,18 +1559,19 @@ void ff_mpv_reconstruct_mb(MpegEncContext *s, int16_t block[12][64])
} }
} }
if (s->encoding) {
mpv_reconstruct_mb_internal(s, block, 0, MAY_BE_MPEG12, 1);
} else if (!s->avctx->lowres) {
#if !CONFIG_SMALL #if !CONFIG_SMALL
if(s->out_format == FMT_MPEG1) { if (s->out_format == FMT_MPEG1)
if (s->avctx->lowres) mpv_reconstruct_mb_internal(s, block, 1, DEFINITELY_MPEG12); mpv_reconstruct_mb_internal(s, block, 0, DEFINITELY_MPEG12, 0);
else mpv_reconstruct_mb_internal(s, block, 0, DEFINITELY_MPEG12); else
} else { mpv_reconstruct_mb_internal(s, block, 0, NOT_MPEG12, 0);
if (s->avctx->lowres) mpv_reconstruct_mb_internal(s, block, 1, NOT_MPEG12);
else mpv_reconstruct_mb_internal(s, block, 0, NOT_MPEG12);
}
#else #else
if (s->avctx->lowres) mpv_reconstruct_mb_internal(s, block, 1, MAY_BE_MPEG12); mpv_reconstruct_mb_internal(s, block, 0, MAY_BE_MPEG12, 0);
else mpv_reconstruct_mb_internal(s, block, 0, MAY_BE_MPEG12);
#endif #endif
} else
mpv_reconstruct_mb_internal(s, block, 1, MAY_BE_MPEG12, 0);
} }
void ff_init_block_index(MpegEncContext *s){ //FIXME maybe rename void ff_init_block_index(MpegEncContext *s){ //FIXME maybe rename