[FFmpeg-devel] [PATCH v2 3/4] avutil/channel_layout: do not copy alloc new map for extra channel layout
Marton Balint
cus at passwd.hu
Wed Mar 16 00:28:58 EET 2022
On Tue, 15 Mar 2022, James Almer wrote:
> On 3/15/2022 6:18 PM, Marton Balint wrote:
>> Also use av_channel_layout_bprint directly for describing channel layout
>> for
>> extra channels.
>>
>> Signed-off-by: Marton Balint <cus at passwd.hu>
>> ---
>> libavutil/channel_layout.c | 17 ++++-------------
>> 1 file changed, 4 insertions(+), 13 deletions(-)
>>
>> diff --git a/libavutil/channel_layout.c b/libavutil/channel_layout.c
>> index 0069c6257b..fb1f72737f 100644
>> --- a/libavutil/channel_layout.c
>> +++ b/libavutil/channel_layout.c
>> @@ -704,29 +704,20 @@ static int try_describe_ambisonic(AVBPrint *bp,
>> const AVChannelLayout *channel_l
>> nb_ambi_channels = (order + 1) * (order + 1);
>> if (nb_ambi_channels < channel_layout->nb_channels) {
>> AVChannelLayout extra = { 0 };
>> - char buf[128];
>>
>> if (channel_layout->order == AV_CHANNEL_ORDER_AMBISONIC) {
>> extra.order = AV_CHANNEL_ORDER_NATIVE;
>> extra.nb_channels = av_popcount64(channel_layout->u.mask);
>> extra.u.mask = channel_layout->u.mask;
>> } else {
>> - const AVChannelCustom *map = channel_layout->u.map;
>> -
>> extra.order = AV_CHANNEL_ORDER_CUSTOM;
>> extra.nb_channels = channel_layout->nb_channels -
>> nb_ambi_channels;
>> - extra.u.map = av_calloc(extra.nb_channels,
>> sizeof(*extra.u.map));
>> - if (!extra.u.map)
>> - return AVERROR(ENOMEM);
>> -
>> - memcpy(extra.u.map, &map[nb_ambi_channels],
>> - sizeof(*extra.u.map) * extra.nb_channels);
>> + extra.u.map = channel_layout->u.map + nb_ambi_channels;
>> }
>> - av_channel_layout_describe(&extra, buf, sizeof(buf));
>> - av_channel_layout_uninit(&extra);
>> -
>> - av_bprintf(bp, "+%s", buf);
>> + av_bprint_chars(bp, '+', 1);
>> + av_channel_layout_describe_bprint(&extra, bp);
>> + /* Not calling uninit here on extra because we don't own the
>> u.map pointer */
>> }
>>
>> return 0;
>
> LGTM.
Thanks, applied the series (although in different order, because it turned
out that this patch depends on the next).
Thanks,
Marton
More information about the ffmpeg-devel
mailing list