On Mon, May 28, 2018 at 01:58:41PM +0300, Laurent Pinchart wrote:
> It would indeed be useful to implement, but I share your concern that this
> would be pretty difficult.
>
> However, there's still something that puzzles me. Let's add a bit more
> context.
>
> if (pipe->num_inputs > 2)
> brx = &vsp1->bru->entity;
> else if (pipe->brx && !drm_pipe->force_brx_release)
> brx = pipe->brx;
> else if (!vsp1->bru->entity.pipe)
> brx = &vsp1->bru->entity;
> else
> brx = &vsp1->brs->entity;
>
> 1. if (!brx)
> return -EINVAL;
>
> 2. if (brx != pipe->brx) {
> ...
> 3. pipe->brx = brx;
> ...
> }
>
> 4. format.pad = pipe->brx->source_pad
>
>
> (1) ensures that brx can't be NULL. (2) is thus always true if pipe->brx is
> NULL. (3) then assigns a non-NULL value to pipe->brx. Smatch should thus never
> complain about (4), even if it can't backtrack.
Oh wow... That's a very basic and ancient bug. I've written a fix and
will push it out tomorrow if it passes my testing.
regards,
dan carpenter
