[FFmpeg-cvslog] avfilter/af_afir: workaround nonsense limitation in vector_fmul_scalar()

[Paul B Mahol]

ffmpeg | branch: master | Paul B Mahol <onemda at gmail.com> | Wed May 10 20:07:30 2017 +0200| [bd404e3949b081788247e2e6e9df0581ef7cc190] | committer: Paul B Mahol

avfilter/af_afir: workaround nonsense limitation in vector_fmul_scalar()

Signed-off-by: Paul B Mahol <onemda at gmail.com>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=bd404e3949b081788247e2e6e9df0581ef7cc190
---

 libavfilter/af_afir.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/libavfilter/af_afir.c b/libavfilter/af_afir.c
index d85c70710e..c4443fdffd 100644
--- a/libavfilter/af_afir.c
+++ b/libavfilter/af_afir.c
@@ -68,7 +68,7 @@ static int fir_channel(AVFilterContext *ctx, void *arg, int ch, int nb_jobs)
     block = s->block[ch] + s->part_index * s->block_size;
     memset(block, 0, sizeof(*block) * s->fft_length);
 
-    s->fdsp->vector_fmul_scalar(block + s->part_size, src, s->dry_gain, s->nb_samples);
+    s->fdsp->vector_fmul_scalar(block + s->part_size, src, s->dry_gain, FFALIGN(s->nb_samples, 4));
     emms_c();
 
     av_rdft_calc(s->rdft[ch], block);
@@ -105,7 +105,7 @@ static int fir_channel(AVFilterContext *ctx, void *arg, int ch, int nb_jobs)
 
     if (out) {
         float *ptr = (float *)out->extended_data[ch];
-        s->fdsp->vector_fmul_scalar(ptr, dst, s->gain * s->wet_gain, out->nb_samples);
+        s->fdsp->vector_fmul_scalar(ptr, dst, s->gain * s->wet_gain, FFALIGN(out->nb_samples, 4));
         emms_c();
     }