[FFmpeg-devel] PATCH: COOK audio decode infastructure to support fixpoint optimization

[Benjamin Larsson]

Marc Hoffman wrote:
> Benjamin Larsson writes:
>  > Marc Hoffman wrote:
>  > > On 7/15/07, Benjamin Larsson <banan at ludd.ltu.se> wrote:
>  > >> Marc Hoffman wrote:
>  > >>
>  > >>> +
>  > >>> +static void joint_decode(COOKContext *q, vec_t* mlt_buffer1,
>  > >>> +                         vec_t* mlt_buffer2) {
>  > >>> +    int decouple_tab[SUBBAND_SIZE];
>  > >>> +    vec_t *decode_buffer;
>  > >>> +
>  > >>> +    decode_buffer = av_malloc (1060*q->f_sample_size);
>  > >> Don't allocate memory with malloc during decode. Use a fixed max size.
>  > > 
>  > > 
>  > > why? your just allocating a really big buffer on the stack which causes
>  > > actually problems for a machine with out a address translation.  Would it be
>  > > acceptable to put that extra memory into COOKContext instead?
>  > > 
>  > > Marc
>  > 
>  > Sure, either preallocate the needed memory or keep it on the stack. The
>  > malloc here could take long time or even fail to allocate the needed memory.
> 
> pre allocate joint_decode/decode_buffer.
> 
> 

Ok.

MvH
Benjamin Larsson