Re: [PATCH] mfreadwrite: Introduce an internal source reader refcount.

[Nikolay Sivov]

On 2/9/22 20:23, Giovanni Mascellani wrote:
> Signed-off-by: Giovanni Mascellani
>
> I think that it might make sense to also check that the media source 
> is released: you keep a reference to it (a reference is taken at the 
> beginning of the test, but immediately released), and check that that 
> too returns zero.

I don't think it should necessarily be zero, it should release what's 
been taken.
> This way you check that the source reader is really released (i.e., 
> there are no other reference loops that prevent 
> src_reader_internal_release() from being called the right number of 
> times).
>
> Thanks, Giovanni.