On Mon, May 28, 2018 at 01:54:18PM +0300, Laurent Pinchart wrote:
> Hi Dan,
>
> Thank you for your quick reply.
>
> On Monday, 28 May 2018 13:20:49 EEST Dan Carpenter wrote:
> > On Mon, May 28, 2018 at 11:28:41AM +0300, Laurent Pinchart wrote:
> > > and I still get the same warning. I had to write the following (which is
> > > obviously not correct) to silence the warning.
> > >
> > > if (pipe->num_inputs > 2)
> > > brx = &vsp1->bru->entity;
> > > else if (pipe->brx)
> > > brx = pipe->brx;
> > > else if (!vsp1->bru->entity.pipe)
> > > brx = &vsp1->bru->entity;
> > > else {
> > > (void)vsp1->brs->entity;
> > > brx = &vsp1->brs->entity;
> > > }
> > >
> > > Both the (void)vsp1->brs->entity and the removal of the !drm_pipe->
> > > force_brx_release were needed, any of those on its own didn't fix the
> > > problem.
> >
> > The problem in this case is the first "brx = &vsp1->bru->entity;".
> > Smatch assumes &vsp1->bru->entity can be == to pipe->brx and NULL.
>
> Why does smatch assume that &vsp1->bru->entity can be NULL, when the previous
> line dereferences vsp1->bru ?
I'm talking about when pipe->num_inputs > 2. For the second
"brx = &vsp1->bru->entity;" assignment, then Smatch parses it correctly
as you say.
regards,
dan carpenter
