Ingo Schwarze <schwa...@usta.de> wrote: > Hi, > > >> A downside of this is that it becomes easier to guess the addresses > >> of the tagged variables. > > > No kidding. It partly undoes the effort of KARL. > > I don't feel qualified to comment on the patch, > but i can't resist mentioning that i still love > tedu@'s classical dictum "attack mitigation countermeasures" > which he coined during the aftermath of the heartbleed debacle.
true, true. That said, it is a small setset of objects. The small set of objects will still be randomly ordered within that region. As the number of objects increases, it becomes harder for an attacker to guess the relative-fixed offset from one object (assuming they get a pointer) to specific other object. So perhaps this is still acceptable?