[FFmpeg-devel] [PATCH 2/3 v3] avcodec/aacdec_template: add more checks to make sure only 22.2 gets to 22.2

[Jan Ekström]

On Sat, Aug 22, 2020 at 11:39 PM Jan Ekström <jeebjp at gmail.com> wrote:
>
> Validates the set channel layout as well as verifies that the received
> layout to the function matches the reference layout, so that it matches
> the implemented re-ordering logic.
>
> Fixes #8845
> ---
>  libavcodec/aacdec_template.c | 21 +++++++++++++++++++--
>  1 file changed, 19 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/aacdec_template.c b/libavcodec/aacdec_template.c
> index 9f7016790e..6473de4ee9 100644
> --- a/libavcodec/aacdec_template.c
> +++ b/libavcodec/aacdec_template.c
> @@ -266,6 +266,7 @@ static int count_paired_channels(uint8_t (*layout_map)[3], int tags, int pos,
>      return num_pos_channels;
>  }
>
> +#define PREFIX_FOR_22POINT2 (AV_CH_LAYOUT_7POINT1_WIDE_BACK|AV_CH_BACK_CENTER|AV_CH_SIDE_LEFT|AV_CH_SIDE_RIGHT|AV_CH_LOW_FREQUENCY_2)
>  static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags)
>  {
>      int i, n, total_non_cc_elements;
> @@ -402,46 +403,60 @@ static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags)
>      }
>
>      // The previous checks would end up at 8 at this point for 22.2
> -    if (tags == 16 && i == 8) {
> +    if (layout == PREFIX_FOR_22POINT2 && tags == 16 && i == 8) {
> +        const uint8_t (*reference_layout_map)[3] = aac_channel_layout_map[12];
> +        for (int j = 0; j < tags; j++) {
> +            if (layout_map[j][0] != reference_layout_map[j][0] ||
> +                layout_map[j][2] != reference_layout_map[j][2])
> +                goto end_of_layout_definition;
> +        }
> +
>          e2c_vec[i] = (struct elem_to_channel) {
>              .av_position  = AV_CH_TOP_FRONT_CENTER,
>              .syn_ele      = layout_map[i][0],
>              .elem_id      = layout_map[i][1],
>              .aac_position = layout_map[i][2]
>          }; layout |= e2c_vec[i].av_position; i++;
> +

These new newlines are extraneous now that the checks before each
assignment are gone.

I don't really care either way at this point, but I did just locally
clean them up - to just make the diff smaller (and to only contain the
bit that matters).

Jan