Message ID | 20181026155606.14754-4-philipl@overt.org |
---|---|
State | Accepted |
Headers | show |
On 26.10.2018 17:56, Philip Langdale wrote: > With the cuda yadif filter in use, the number of mapped decoder > frames could increase by two, as the filter holds on to additional > frames. > > Signed-off-by: Philip Langdale <philipl@overt.org> > --- > libavcodec/nvdec.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/libavcodec/nvdec.c b/libavcodec/nvdec.c > index 4dd6b1acf3..8e91c36084 100644 > --- a/libavcodec/nvdec.c > +++ b/libavcodec/nvdec.c > @@ -601,7 +601,7 @@ int ff_nvdec_frame_params(AVCodecContext *avctx, > frames_ctx->format = AV_PIX_FMT_CUDA; > frames_ctx->width = (avctx->coded_width + 1) & ~1; > frames_ctx->height = (avctx->coded_height + 1) & ~1; > - frames_ctx->initial_pool_size = dpb_size; > + frames_ctx->initial_pool_size = dpb_size + 2; A small comment here as to where that magic +2 is coming from and it's GTM.
diff --git a/libavcodec/nvdec.c b/libavcodec/nvdec.c index 4dd6b1acf3..8e91c36084 100644 --- a/libavcodec/nvdec.c +++ b/libavcodec/nvdec.c @@ -601,7 +601,7 @@ int ff_nvdec_frame_params(AVCodecContext *avctx, frames_ctx->format = AV_PIX_FMT_CUDA; frames_ctx->width = (avctx->coded_width + 1) & ~1; frames_ctx->height = (avctx->coded_height + 1) & ~1; - frames_ctx->initial_pool_size = dpb_size; + frames_ctx->initial_pool_size = dpb_size + 2; frames_ctx->free = nvdec_free_dummy; frames_ctx->pool = av_buffer_pool_init(0, nvdec_alloc_dummy);
With the cuda yadif filter in use, the number of mapped decoder frames could increase by two, as the filter holds on to additional frames. Signed-off-by: Philip Langdale <philipl@overt.org> --- libavcodec/nvdec.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)