> For example Astrée does not support dynamic memory allocation nor recursion (not too bad for embedded). Are you sure? Around 2019, they seem to have overcome

> For example Astrée does not support dynamic memory allocation nor recursion (not too bad for embedded). Are you sure? Around 2019, they seem to have overcome those limitations, since they removed mention of them from their website. Another OpenZFS developer found that on an old university page and pointed it out to me, but that webpage was made well before then. I wonder if you read the same page that he did. I plan to look into that when I am using the free trial. > To do better, a fully automatic sound analyzer would have to derive the relevant loop invariants and the associated induction proofs automatically, not to have to make any guess on the number of iterations. As far as I know this is still a research topic. Or alternatively, enforce a known, small enough bounds for any loop. I plan to look into this when I try using sound static analyzers, since if they cannot do induction proofs on loops to prove loop correctness, it would either invalidate their claims of soundness, or at the very least limit the scope of such claims, which would mean that I would not be able to semi-formally verify ZFS using them. I know that frama-c’s Eva is specifically documented as not being able to do this.