[FFmpeg-devel] [PATCH v2] avutil/frame: add av_frame_replace
James Almer
jamrial at gmail.com
Tue Jun 7 14:39:55 EEST 2022
On 6/7/2022 7:21 AM, Anton Khirnov wrote:
> Quoting James Almer (2022-06-05 19:44:58)
>> +
>> + /* duplicate the frame data if it's not refcounted */
>> + if (!src->buf[0]) {
>> + for (int i = 0; i < FF_ARRAY_ELEMS(dst->buf); i++)
>> + av_buffer_unref(&dst->buf[i]);
>> + for (int i = 0; i < dst->nb_extended_buf; i++)
>> + av_buffer_unref(&dst->extended_buf[i]);
>> + av_freep(&dst->extended_buf);
>> +
>> + memset(dst->data, 0, sizeof(dst->data));
>> + if (dst->extended_data != dst->data)
>> + av_freep(&dst->extended_data);
>> +
>> + ret = av_frame_get_buffer(dst, 0);
>> + if (ret < 0)
>> + goto fail;
>> +
>> + ret = av_frame_copy(dst, src);
>> + if (ret < 0)
>> + goto fail;
>> +
>> + ret = av_buffer_replace(&dst->hw_frames_ctx, src->hw_frames_ctx);
>> + if (ret < 0)
>> + goto fail;
>
> This looks suspicious, since we just allocated normal buffers for dst.
Right, removed.
> Also, can't t this whole block be replaced by just
>
> if (!src->buf[0]) {
> av_frame_unref(dst);
> return av_frame_ref(dst, src);
> }
>
> at the beginning of this function?
Yes, but i was copying how av_frame_ref() handled this scenario, albeit
with a bit more of code.
Will change, but it will need to be re-added if someone at some point
decides to implement replace for side data.
>
>> + if (src->extended_data != src->data) {
>> + int ch = src->ch_layout.nb_channels;
>> +
>> +#if FF_API_OLD_CHANNEL_LAYOUT
>> +FF_DISABLE_DEPRECATION_WARNINGS
>> + if (!ch) {
>> + ch = src->channels;
>> + CHECK_CHANNELS_CONSISTENCY(src);
>> + }
>> +FF_ENABLE_DEPRECATION_WARNINGS
>> +#endif
>> + if (!ch) {
>> + ret = AVERROR(EINVAL);
>> + goto fail;
>> + }
>> +
>> + dst->extended_data = av_malloc_array(sizeof(*dst->extended_data), ch);
>> + if (!dst->extended_data) {
>> + ret = AVERROR(ENOMEM);
>> + goto fail;
>> + }
>> + memcpy(dst->extended_data, src->extended_data, sizeof(*src->extended_data) * ch);
>
> nit: av_memdup()?
>
>> + } else
>> + dst->extended_data = dst->data;
>> +
>> + memcpy(dst->data, src->data, sizeof(src->data));
>> + memcpy(dst->linesize, src->linesize, sizeof(src->linesize));
>> +
>> + return 0;
>> +
>> +fail:
>> + av_frame_unref(dst);
>> + return ret;
>> +}
>> +
>> AVFrame *av_frame_clone(const AVFrame *src)
>> {
>> AVFrame *ret = av_frame_alloc();
>> diff --git a/libavutil/frame.h b/libavutil/frame.h
>> index 33fac2054c..e5c10e2b66 100644
>> --- a/libavutil/frame.h
>> +++ b/libavutil/frame.h
>> @@ -752,6 +752,19 @@ void av_frame_free(AVFrame **frame);
>> */
>> int av_frame_ref(AVFrame *dst, const AVFrame *src);
>>
>> +/**
>> + * Ensure the destination frame refers to the same data described by the source
>> + * frame by creating a new reference for each AVBufferRef from src if they
>> + * differ from those in dst, or if src is not reference counted, by allocating
>> + * new buffers and copying data.
>> + *
>> + * Frame properties on dst will be replaced by those from src.
>> + *
>> + * @return 0 on success, a negative AVERROR on error. On error, dst is
>> + * unreferenced.
>> + */
>> +int av_frame_replace(AVFrame *dst, const AVFrame *src);
>
> An important property of av_buffer_replace() is that it's equivalent to
> av_buffer_unref(dst) when src is NULL. It would probably be desirable
> for av_frame_replace() to work the same way - currently it will try to
> call av_frame_get_buffer() with invalid parameters and fail.
Should it really accept NULL as src, or you meant a recently
allocated/unreff'd frame as src (So src->data[0] == NULL)?
More information about the ffmpeg-devel
mailing list