> > Convert the route expire timestamp in kernel and routing message > > to 64 bit. Add a small compatibility layer that allows to set > > routes with old user land and new kernel. > > > > ok? > > just out for curiosity, shouldn't the expire time be u_int64_t?
It seems such a simple concept. It should never be signed, so let's just treat it as unsigned. Except if it goes unsigned, the same problems develop as if when it is signed. So the rule remains the same. This is the same mess that happened with ssize_t and size_t. A better rule would have been to restrict both sizes to the same limit. But no..... we assume people can be careful, or the little 'u' which avoids a warning will make things better