Browse Source

avcodec/scpr: Use av_memcpy_backptr() in type 17 and 33

This makes the changed code-path faster.

Change not tested except with the fuzzer testcase as I found no other testcase.

Improves: Timeout (136sec -> 74sec)
Improves: 16040/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SCPR_fuzzer-5705876062601216

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
(cherry picked from commit 950a21e83c)
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
tags/n4.0.5
Michael Niedermayer 6 years ago
parent
commit
2ce3997846
1 changed files with 3 additions and 4 deletions
  1. +3
    -4
      libavcodec/scpr.c

+ 3
- 4
libavcodec/scpr.c View File

@@ -786,7 +786,7 @@ static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
s->current_frame->linesize[0] / 4);
} else if (type == 17) {
uint32_t clr, *dst = (uint32_t *)s->current_frame->data[0];
int x, y;
int y;

frame->key_frame = 1;
bytestream2_skip(gb, 1);
@@ -802,9 +802,8 @@ static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
clr = bytestream2_get_le24(gb);
}
for (y = 0; y < avctx->height; y++) {
for (x = 0; x < avctx->width; x++) {
dst[x] = clr;
}
dst[0] = clr;
av_memcpy_backptr((uint8_t*)(dst+1), 4, 4*avctx->width - 4);
dst += s->current_frame->linesize[0] / 4;
}
} else if (type == 0 || type == 1) {


Loading…
Cancel
Save