ffmpeg/libavcodec/vp9_mc_template.c
Andreas Rheinhardt 7bd3b73716 avcodec/vp9: Switch to ProgressFrames
This already fixes a race in the vp9-encparams test. In this test,
side data is added to the current frame after having been decoded
(and therefore after ff_thread_finish_setup() has been called).
Yet the update_thread_context callback called ff_thread_ref_frame()
and therefore av_frame_ref() with this frame as source frame and
the ensuing read was unsynchronised with adding the side data,
i.e. there was a data race.

By switching to the ProgressFrame API the implicit av_frame_ref()
is removed and the race fixed except if this frame is later reused by
a show-existing-frame which uses an explicit av_frame_ref().
The vp9-encparams test does not cover this, so this commit
already fixes all the races in this test.

This decoder kept multiple references to the same ThreadFrames
in the same context and therefore had lots of implicit av_frame_ref()
even when decoding single-threaded. This incurred lots of small
allocations: When decoding an ordinary 10s video in single-threaded
mode the number of allocations reported by Valgrind went down
from 57,814 to 20,908; for 10 threads it went down from 84,223 to
21,901.

Reviewed-by: Anton Khirnov <anton@khirnov.net>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-04-19 13:18:04 +02:00

440 lines
24 KiB
C

/*
* VP9 compatible video decoder
*
* Copyright (C) 2013 Ronald S. Bultje <rsbultje gmail com>
* Copyright (C) 2013 Clément Bœsch <u pkh me>
*
* 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
*/
#define ROUNDED_DIV_MVx2(a, b) \
(VP9mv) { .x = ROUNDED_DIV(a.x + b.x, 2), .y = ROUNDED_DIV(a.y + b.y, 2) }
#define ROUNDED_DIV_MVx4(a, b, c, d) \
(VP9mv) { .x = ROUNDED_DIV(a.x + b.x + c.x + d.x, 4), \
.y = ROUNDED_DIV(a.y + b.y + c.y + d.y, 4) }
static void FN(inter_pred)(VP9TileData *td)
{
static const uint8_t bwlog_tab[2][N_BS_SIZES] = {
{ 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4 },
{ 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 4, 4 },
};
const VP9Context *s = td->s;
VP9Block *b = td->b;
int row = td->row, col = td->col;
const ProgressFrame *tref1 = &s->s.refs[s->s.h.refidx[b->ref[0]]], *tref2;
const AVFrame *ref1 = tref1->f, *ref2;
int w1 = ref1->width, h1 = ref1->height, w2, h2;
ptrdiff_t ls_y = td->y_stride, ls_uv = td->uv_stride;
int bytesperpixel = BYTES_PER_PIXEL;
if (b->comp) {
tref2 = &s->s.refs[s->s.h.refidx[b->ref[1]]];
ref2 = tref2->f;
w2 = ref2->width;
h2 = ref2->height;
}
// y inter pred
if (b->bs > BS_8x8) {
VP9mv uvmv;
#if SCALED == 0
if (b->bs == BS_8x4) {
mc_luma_dir(td, mc[3][b->filter][0], td->dst[0], ls_y,
ref1->data[0], ref1->linesize[0], tref1,
row << 3, col << 3, &b->mv[0][0],,,,, 8, 4, w1, h1, 0);
mc_luma_dir(td, mc[3][b->filter][0],
td->dst[0] + 4 * ls_y, ls_y,
ref1->data[0], ref1->linesize[0], tref1,
(row << 3) + 4, col << 3, &b->mv[2][0],,,,, 8, 4, w1, h1, 0);
w1 = (w1 + s->ss_h) >> s->ss_h;
if (s->ss_v) {
h1 = (h1 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx2(b->mv[0][0], b->mv[2][0]);
mc_chroma_dir(td, mc[3 + s->ss_h][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 2, col << (3 - s->ss_h),
&uvmv,,,,, 8 >> s->ss_h, 4, w1, h1, 0);
} else {
mc_chroma_dir(td, mc[3 + s->ss_h][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 3, col << (3 - s->ss_h),
&b->mv[0][0],,,,, 8 >> s->ss_h, 4, w1, h1, 0);
// BUG for 4:2:2 bs=8x4, libvpx uses the wrong block index
// to get the motion vector for the bottom 4x4 block
// https://code.google.com/p/webm/issues/detail?id=993
if (s->ss_h == 0) {
uvmv = b->mv[2][0];
} else {
uvmv = ROUNDED_DIV_MVx2(b->mv[0][0], b->mv[2][0]);
}
mc_chroma_dir(td, mc[3 + s->ss_h][b->filter][0],
td->dst[1] + 4 * ls_uv, td->dst[2] + 4 * ls_uv, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
(row << 3) + 4, col << (3 - s->ss_h),
&uvmv,,,,, 8 >> s->ss_h, 4, w1, h1, 0);
}
if (b->comp) {
mc_luma_dir(td, mc[3][b->filter][1], td->dst[0], ls_y,
ref2->data[0], ref2->linesize[0], tref2,
row << 3, col << 3, &b->mv[0][1],,,,, 8, 4, w2, h2, 1);
mc_luma_dir(td, mc[3][b->filter][1],
td->dst[0] + 4 * ls_y, ls_y,
ref2->data[0], ref2->linesize[0], tref2,
(row << 3) + 4, col << 3, &b->mv[2][1],,,,, 8, 4, w2, h2, 1);
w2 = (w2 + s->ss_h) >> s->ss_h;
if (s->ss_v) {
h2 = (h2 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx2(b->mv[0][1], b->mv[2][1]);
mc_chroma_dir(td, mc[3 + s->ss_h][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 2, col << (3 - s->ss_h),
&uvmv,,,,, 8 >> s->ss_h, 4, w2, h2, 1);
} else {
mc_chroma_dir(td, mc[3 + s->ss_h][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 3, col << (3 - s->ss_h),
&b->mv[0][1],,,,, 8 >> s->ss_h, 4, w2, h2, 1);
// BUG for 4:2:2 bs=8x4, libvpx uses the wrong block index
// to get the motion vector for the bottom 4x4 block
// https://code.google.com/p/webm/issues/detail?id=993
if (s->ss_h == 0) {
uvmv = b->mv[2][1];
} else {
uvmv = ROUNDED_DIV_MVx2(b->mv[0][1], b->mv[2][1]);
}
mc_chroma_dir(td, mc[3 + s->ss_h][b->filter][1],
td->dst[1] + 4 * ls_uv, td->dst[2] + 4 * ls_uv, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
(row << 3) + 4, col << (3 - s->ss_h),
&uvmv,,,,, 8 >> s->ss_h, 4, w2, h2, 1);
}
}
} else if (b->bs == BS_4x8) {
mc_luma_dir(td, mc[4][b->filter][0], td->dst[0], ls_y,
ref1->data[0], ref1->linesize[0], tref1,
row << 3, col << 3, &b->mv[0][0],,,,, 4, 8, w1, h1, 0);
mc_luma_dir(td, mc[4][b->filter][0], td->dst[0] + 4 * bytesperpixel, ls_y,
ref1->data[0], ref1->linesize[0], tref1,
row << 3, (col << 3) + 4, &b->mv[1][0],,,,, 4, 8, w1, h1, 0);
h1 = (h1 + s->ss_v) >> s->ss_v;
if (s->ss_h) {
w1 = (w1 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx2(b->mv[0][0], b->mv[1][0]);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << (3 - s->ss_v), col << 2,
&uvmv,,,,, 4, 8 >> s->ss_v, w1, h1, 0);
} else {
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << (3 - s->ss_v), col << 3,
&b->mv[0][0],,,,, 4, 8 >> s->ss_v, w1, h1, 0);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1] + 4 * bytesperpixel,
td->dst[2] + 4 * bytesperpixel, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << (3 - s->ss_v), (col << 3) + 4,
&b->mv[1][0],,,,, 4, 8 >> s->ss_v, w1, h1, 0);
}
if (b->comp) {
mc_luma_dir(td, mc[4][b->filter][1], td->dst[0], ls_y,
ref2->data[0], ref2->linesize[0], tref2,
row << 3, col << 3, &b->mv[0][1],,,,, 4, 8, w2, h2, 1);
mc_luma_dir(td, mc[4][b->filter][1], td->dst[0] + 4 * bytesperpixel, ls_y,
ref2->data[0], ref2->linesize[0], tref2,
row << 3, (col << 3) + 4, &b->mv[1][1],,,,, 4, 8, w2, h2, 1);
h2 = (h2 + s->ss_v) >> s->ss_v;
if (s->ss_h) {
w2 = (w2 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx2(b->mv[0][1], b->mv[1][1]);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << (3 - s->ss_v), col << 2,
&uvmv,,,,, 4, 8 >> s->ss_v, w2, h2, 1);
} else {
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << (3 - s->ss_v), col << 3,
&b->mv[0][1],,,,, 4, 8 >> s->ss_v, w2, h2, 1);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1] + 4 * bytesperpixel,
td->dst[2] + 4 * bytesperpixel, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << (3 - s->ss_v), (col << 3) + 4,
&b->mv[1][1],,,,, 4, 8 >> s->ss_v, w2, h2, 1);
}
}
} else
#endif
{
#if SCALED == 0
av_assert2(b->bs == BS_4x4);
#endif
// FIXME if two horizontally adjacent blocks have the same MV,
// do a w8 instead of a w4 call
mc_luma_dir(td, mc[4][b->filter][0], td->dst[0], ls_y,
ref1->data[0], ref1->linesize[0], tref1,
row << 3, col << 3, &b->mv[0][0],
0, 0, 8, 8, 4, 4, w1, h1, 0);
mc_luma_dir(td, mc[4][b->filter][0], td->dst[0] + 4 * bytesperpixel, ls_y,
ref1->data[0], ref1->linesize[0], tref1,
row << 3, (col << 3) + 4, &b->mv[1][0],
4, 0, 8, 8, 4, 4, w1, h1, 0);
mc_luma_dir(td, mc[4][b->filter][0],
td->dst[0] + 4 * ls_y, ls_y,
ref1->data[0], ref1->linesize[0], tref1,
(row << 3) + 4, col << 3, &b->mv[2][0],
0, 4, 8, 8, 4, 4, w1, h1, 0);
mc_luma_dir(td, mc[4][b->filter][0],
td->dst[0] + 4 * ls_y + 4 * bytesperpixel, ls_y,
ref1->data[0], ref1->linesize[0], tref1,
(row << 3) + 4, (col << 3) + 4, &b->mv[3][0],
4, 4, 8, 8, 4, 4, w1, h1, 0);
if (s->ss_v) {
h1 = (h1 + 1) >> 1;
if (s->ss_h) {
w1 = (w1 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx4(b->mv[0][0], b->mv[1][0],
b->mv[2][0], b->mv[3][0]);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 2, col << 2,
&uvmv, 0, 0, 4, 4, 4, 4, w1, h1, 0);
} else {
uvmv = ROUNDED_DIV_MVx2(b->mv[0][0], b->mv[2][0]);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 2, col << 3,
&uvmv, 0, 0, 8, 4, 4, 4, w1, h1, 0);
uvmv = ROUNDED_DIV_MVx2(b->mv[1][0], b->mv[3][0]);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1] + 4 * bytesperpixel,
td->dst[2] + 4 * bytesperpixel, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 2, (col << 3) + 4,
&uvmv, 4, 0, 8, 4, 4, 4, w1, h1, 0);
}
} else {
if (s->ss_h) {
w1 = (w1 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx2(b->mv[0][0], b->mv[1][0]);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 3, col << 2,
&uvmv, 0, 0, 4, 8, 4, 4, w1, h1, 0);
// BUG libvpx uses wrong block index for 4:2:2 bs=4x4
// bottom block
// https://code.google.com/p/webm/issues/detail?id=993
uvmv = ROUNDED_DIV_MVx2(b->mv[1][0], b->mv[2][0]);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1] + 4 * ls_uv, td->dst[2] + 4 * ls_uv, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
(row << 3) + 4, col << 2,
&uvmv, 0, 4, 4, 8, 4, 4, w1, h1, 0);
} else {
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 3, col << 3,
&b->mv[0][0], 0, 0, 8, 8, 4, 4, w1, h1, 0);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1] + 4 * bytesperpixel,
td->dst[2] + 4 * bytesperpixel, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << 3, (col << 3) + 4,
&b->mv[1][0], 4, 0, 8, 8, 4, 4, w1, h1, 0);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1] + 4 * ls_uv, td->dst[2] + 4 * ls_uv, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
(row << 3) + 4, col << 3,
&b->mv[2][0], 0, 4, 8, 8, 4, 4, w1, h1, 0);
mc_chroma_dir(td, mc[4][b->filter][0],
td->dst[1] + 4 * ls_uv + 4 * bytesperpixel,
td->dst[2] + 4 * ls_uv + 4 * bytesperpixel, ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
(row << 3) + 4, (col << 3) + 4,
&b->mv[3][0], 4, 4, 8, 8, 4, 4, w1, h1, 0);
}
}
if (b->comp) {
mc_luma_dir(td, mc[4][b->filter][1], td->dst[0], ls_y,
ref2->data[0], ref2->linesize[0], tref2,
row << 3, col << 3, &b->mv[0][1], 0, 0, 8, 8, 4, 4, w2, h2, 1);
mc_luma_dir(td, mc[4][b->filter][1], td->dst[0] + 4 * bytesperpixel, ls_y,
ref2->data[0], ref2->linesize[0], tref2,
row << 3, (col << 3) + 4, &b->mv[1][1], 4, 0, 8, 8, 4, 4, w2, h2, 1);
mc_luma_dir(td, mc[4][b->filter][1],
td->dst[0] + 4 * ls_y, ls_y,
ref2->data[0], ref2->linesize[0], tref2,
(row << 3) + 4, col << 3, &b->mv[2][1], 0, 4, 8, 8, 4, 4, w2, h2, 1);
mc_luma_dir(td, mc[4][b->filter][1],
td->dst[0] + 4 * ls_y + 4 * bytesperpixel, ls_y,
ref2->data[0], ref2->linesize[0], tref2,
(row << 3) + 4, (col << 3) + 4, &b->mv[3][1], 4, 4, 8, 8, 4, 4, w2, h2, 1);
if (s->ss_v) {
h2 = (h2 + 1) >> 1;
if (s->ss_h) {
w2 = (w2 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx4(b->mv[0][1], b->mv[1][1],
b->mv[2][1], b->mv[3][1]);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 2, col << 2,
&uvmv, 0, 0, 4, 4, 4, 4, w2, h2, 1);
} else {
uvmv = ROUNDED_DIV_MVx2(b->mv[0][1], b->mv[2][1]);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 2, col << 3,
&uvmv, 0, 0, 8, 4, 4, 4, w2, h2, 1);
uvmv = ROUNDED_DIV_MVx2(b->mv[1][1], b->mv[3][1]);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1] + 4 * bytesperpixel,
td->dst[2] + 4 * bytesperpixel, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 2, (col << 3) + 4,
&uvmv, 4, 0, 8, 4, 4, 4, w2, h2, 1);
}
} else {
if (s->ss_h) {
w2 = (w2 + 1) >> 1;
uvmv = ROUNDED_DIV_MVx2(b->mv[0][1], b->mv[1][1]);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 3, col << 2,
&uvmv, 0, 0, 4, 8, 4, 4, w2, h2, 1);
// BUG libvpx uses wrong block index for 4:2:2 bs=4x4
// bottom block
// https://code.google.com/p/webm/issues/detail?id=993
uvmv = ROUNDED_DIV_MVx2(b->mv[1][1], b->mv[2][1]);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1] + 4 * ls_uv, td->dst[2] + 4 * ls_uv, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
(row << 3) + 4, col << 2,
&uvmv, 0, 4, 4, 8, 4, 4, w2, h2, 1);
} else {
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 3, col << 3,
&b->mv[0][1], 0, 0, 8, 8, 4, 4, w2, h2, 1);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1] + 4 * bytesperpixel,
td->dst[2] + 4 * bytesperpixel, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << 3, (col << 3) + 4,
&b->mv[1][1], 4, 0, 8, 8, 4, 4, w2, h2, 1);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1] + 4 * ls_uv, td->dst[2] + 4 * ls_uv, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
(row << 3) + 4, col << 3,
&b->mv[2][1], 0, 4, 8, 8, 4, 4, w2, h2, 1);
mc_chroma_dir(td, mc[4][b->filter][1],
td->dst[1] + 4 * ls_uv + 4 * bytesperpixel,
td->dst[2] + 4 * ls_uv + 4 * bytesperpixel, ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
(row << 3) + 4, (col << 3) + 4,
&b->mv[3][1], 4, 4, 8, 8, 4, 4, w2, h2, 1);
}
}
}
}
} else {
int bwl = bwlog_tab[0][b->bs];
int bw = ff_vp9_bwh_tab[0][b->bs][0] * 4;
int bh = ff_vp9_bwh_tab[0][b->bs][1] * 4;
int uvbw = ff_vp9_bwh_tab[s->ss_h][b->bs][0] * 4;
int uvbh = ff_vp9_bwh_tab[s->ss_v][b->bs][1] * 4;
mc_luma_dir(td, mc[bwl][b->filter][0], td->dst[0], ls_y,
ref1->data[0], ref1->linesize[0], tref1,
row << 3, col << 3, &b->mv[0][0], 0, 0, bw, bh, bw, bh, w1, h1, 0);
w1 = (w1 + s->ss_h) >> s->ss_h;
h1 = (h1 + s->ss_v) >> s->ss_v;
mc_chroma_dir(td, mc[bwl + s->ss_h][b->filter][0],
td->dst[1], td->dst[2], ls_uv,
ref1->data[1], ref1->linesize[1],
ref1->data[2], ref1->linesize[2], tref1,
row << (3 - s->ss_v), col << (3 - s->ss_h),
&b->mv[0][0], 0, 0, uvbw, uvbh, uvbw, uvbh, w1, h1, 0);
if (b->comp) {
mc_luma_dir(td, mc[bwl][b->filter][1], td->dst[0], ls_y,
ref2->data[0], ref2->linesize[0], tref2,
row << 3, col << 3, &b->mv[0][1], 0, 0, bw, bh, bw, bh, w2, h2, 1);
w2 = (w2 + s->ss_h) >> s->ss_h;
h2 = (h2 + s->ss_v) >> s->ss_v;
mc_chroma_dir(td, mc[bwl + s->ss_h][b->filter][1],
td->dst[1], td->dst[2], ls_uv,
ref2->data[1], ref2->linesize[1],
ref2->data[2], ref2->linesize[2], tref2,
row << (3 - s->ss_v), col << (3 - s->ss_h),
&b->mv[0][1], 0, 0, uvbw, uvbh, uvbw, uvbh, w2, h2, 1);
}
}
}