News for hackers. Stuff that matters.
Re:/dev/random – a new approach by acqq on Sunday April 24, 2016 @06:27PM (#11561654)
If I understood it, this change is supposed to be fully compatible with the existing interfaces and assumptions, it just claims to give more bits faster, using Fortuna would mean more changes at once and would mean the client code would have to change. Which is easier for OpenBSD to do than it is for Linux.
Re:/dev/random – a new approach by tptacek on Sunday April 24, 2016 @07:38PM (#11561845)
Replacing the LRNG with Yarrow or Fortuna or something like it would not break any client code, because no sane client code can rely on when /dev/random decides to block --- even by the LRNG's broken definition, that blocking happens at random, and could be deferred indefinitely if enough "entropy" were available.
Re:/dev/random – a new approach by twic on Sunday April 24, 2016 @06:46PM (#11561714)
How would it change the client code? If you use Fortuna to implement /dev/{u,}random, then the client code can just carry on opening that and reading as many bytes as it needs.