[FFmpeg-devel] [PATCH 1/2] Revert "avfilter/vf_minterpolate: if metadata lavfi.scd.mafd exists, we'll use it first"
Paul B Mahol
onemda at gmail.com
Fri May 15 19:42:59 EEST 2020
Why was this applied without review?
Please revert ASAP!
On 5/14/20, Marton Balint <cus at passwd.hu> wrote:
> This reverts commit d88e1c9838dbcfe29d7835f2705ffc9ee6a36bf3.
>
> Fixes null pointer dereference.
>
> Signed-off-by: Marton Balint <cus at passwd.hu>
> ---
> libavfilter/vf_minterpolate.c | 16 ++++------------
> 1 file changed, 4 insertions(+), 12 deletions(-)
>
> diff --git a/libavfilter/vf_minterpolate.c b/libavfilter/vf_minterpolate.c
> index ef47140946..c9ce80420d 100644
> --- a/libavfilter/vf_minterpolate.c
> +++ b/libavfilter/vf_minterpolate.c
> @@ -834,19 +834,11 @@ static int detect_scene_change(MIContext *mi_ctx)
> ptrdiff_t linesize2 = mi_ctx->frames[2].avf->linesize[0];
>
> if (mi_ctx->scd_method == SCD_METHOD_FDIFF) {
> - double ret = 0, mafd = HUGE_VAL, diff;
> + double ret = 0, mafd, diff;
> uint64_t sad;
> - AVDictionaryEntry *e_mafd = NULL;
> - char *tail = NULL;
> -
> - e_mafd = av_dict_get(mi_ctx->frames[2].avf->metadata,
> "lavfi.scd.mafd", NULL, AV_DICT_MATCH_CASE);
> - if (e_mafd)
> - mafd = strtod(e_mafd->value, &tail);
> - if (*tail || mafd == HUGE_VAL) {
> - mi_ctx->sad(p1, linesize1, p2, linesize2, me_ctx->width,
> me_ctx->height, &sad);
> - emms_c();
> - mafd = (double) sad * 100.0 / (me_ctx->height * me_ctx->width)
> / (1 << mi_ctx->bitdepth);
> - }
> + mi_ctx->sad(p1, linesize1, p2, linesize2, me_ctx->width,
> me_ctx->height, &sad);
> + emms_c();
> + mafd = (double) sad * 100.0 / (me_ctx->height * me_ctx->width) / (1
> << mi_ctx->bitdepth);
> diff = fabs(mafd - mi_ctx->prev_mafd);
> ret = av_clipf(FFMIN(mafd, diff), 0, 100.0);
> mi_ctx->prev_mafd = mafd;
> --
> 2.16.4
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
More information about the ffmpeg-devel
mailing list