search

secure-software-engineering / phasar

A LLVM-based static analysis framework.
Other
919 stars 140 forks source link

Fix handling of unbalanced returns for IIA #664

Closed MMory closed 10 months ago

MMory commented 10 months ago

For the change in the IDE solver I am not sure yet how this should be done properly, let's discuss that @fabianbs96

fabianbs96 commented 10 months ago

@MMory I suppose this is correct; otherwise we cannot propagate back seed parameters in an unbalanced return, however, I don't know the exact reason for this being introduced (this is already there in Heros). On the other hand, for what use case do you need this?

Regarding the null-check on CallSite in the ret-FF, I don't think, this is the way to go. We should rather change the one location in the IDESolver where the call-site is passed nullptr an extra API in the IDETabulationProblem that can be overridden if required (I think, this is required only very rarely). Further, as the IDESolver is generic over the to be analyzed IR, we should also be able to handle statement (n_t) types that are not nullable.

EDIT: The second part may be a separate PR, or what do you think?