[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Not enough entropy in RNG



On Tue, Jan 29, 2002 at 10:29:03AM -0500, Derek L Davies wrote:
> 
> Thanks -- I've download and untarred it.  It's, er, slightly short on
> instructions ;-)  I see an executable named 'random' do I just copy
> that over to my /dev directory or something?  How do I get urandom?

You'll have to search for my announcement on one of the lists ;)

Try random --help, and use it as a translator, like this:

settrans -c /dev/random /hurd/random

urandom requires some option to make it not block, I forgot all the details.
 
> I don't know much about gathering entropy, other than my memory is
> pretty much shot these days (drum hit).  When you say "feed it" I assume
> you mean just use the system.

My random translator is fed by writing entropy to /dev/random.  The entropy
is raw bytes, not qualified (it is assumed that one byte entropy is good to
generate one byte random).  I have surely posted some notes on what to do to
get system entropy in this translator into my announcement.

In particular, the random translator should use poll mode rather than wait
for entropy.  This requires a kernel device, and a libchannel abstraction
would also be nice to have.

But the main problem is entropy of course.  Ifyou have installed the
translator, try "cat /dev/random" in one terminal and
"cat /myrandombytes > /dev/random" in another, where /myrandombytes contains
good randomness.

> Searching the web I saw stuff about
> daemon programs that one can run to generate entropy.  I wouldn't mind
> trying to get one of these going if thats what you're alluding to.  Is
> there a specific program that I should start with?

The perl entrophy generator daemon or so (egd).

> (Sorry if this is documented.  I read all the stuff I could find.  I
> can write up a quick INSTALL covering what I learn about this that
> Marcus could include in the random tarball if people would like.)

Well, it is just a prototype.  Eventually all this stuff will end up in the
distro, with entries in MAKEDEV, docs in hurd.texi, and so on.  The details
are subject to changes, so I'd prefer it if you put your energy into
improving the code to cover the missing things ;)

Thanks,
Marcus

-- 
`Rhubarb is no Egyptian god.' Debian http://www.debian.org brinkmd@debian.org
Marcus Brinkmann              GNU    http://www.gnu.org    marcus@gnu.org
Marcus.Brinkmann@ruhr-uni-bochum.de
http://www.marcus-brinkmann.de



Reply to: