On Wed, 30 Oct 2019, Markus Elfring wrote:

> > There is no reason why a patch should be generated in this case.
> > As you should know well, A ... B only matches in a transformation case
> > if every path from A leads to code matching B.  That is not the case in your example.
>
> The exception handling code should usually be executed at the end of
> function implementations after an error situation was detected.
>
> Do you try to refer to specific information from the software documentation
> like the following?
>
> https://github.com/coccinelle/coccinelle/blob/ed1eb8e06f800739d3992158d36945c0c4c6f0c7/docs/manual/cocci_syntax.tex#L179
> “…
> A depends on clause can further indicate whether the clause should be satisfied
> for all the branches (forall) or only for one (exists). exists is the default.
> …”

There is no depends on in your code.  The cited text is not relevant.

julia

>
> The following simple transformation approach seems to work in the way
> which I expected somehow initially.
>
> @addition exists@
> expression object;
> @@
>  object = kzalloc(...)
>  ... when any
>  device_unregister(...);
>  out:
> +kfree(object);
>  return ERR_PTR(...);
>
>
> Does this change specification indicate then a disagreement about
> a default SmPL rule property?
>
> Regards,
> Markus
>