Follow-up Comment #1, task #5130 (project hurd):
I figure this is a good place to post. For anyone following IRC, I've been
fighting to get a working entropy device in the kernel, and finally managed
to get some entropy going (not much, but its only reading one source as of
writing, I'll fix that tommorow). As it stands, using storeio with the
entropy device creates a perfectly usable /dev/random, but its going to
likely need a custom translator to create /dev/urandom.
I'm also not sure if I should entropy just dump raw bits, and let the
translator mash it up, or have it happen in the kernel itself. I'm leaning
towards doing the actual mashing within the kernel, and have an option to
turn that on and off. Anyone have any opinions on the subject or not.
I'm also trying to figure out how to handle entropy collection for urandom.
What needed is the buffer to change constantly even if new seed data can't be
gathered, which is what Linux does, and I suggest we simply use the same
agrothim as them since we don't need to prove ours is then secure.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/task/?5130>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
_______________________________________________
Bug-hurd mailing list
Bug-hurd@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-hurd
