Re: [Tails-dev] How to seed urandom (or not)?

Borrar esta mensaxe

Responder a esta mensaxe
Autor: coderman
Data:  
Para: David Goulet
CC: The Tails public development discussion list
Asunto: Re: [Tails-dev] How to seed urandom (or not)?
On Sun, Aug 3, 2014 at 12:35 PM, David Goulet <dgoulet@???> wrote:
> ...
> I asked a friend of mine working for Canonical asking him how they do
> things with Ubuntu images.
>
> He told me that they can't use haveged since it's only good to
> *increase* your entropy pool and rngd is useless without an hardware
> RNG. Their big issue is the Ubuntu Cloud Image for which they rely on
> https://launchpad.net/pollinate,


yes; the common "worst case scenario" along with embedded devices. i
didn't mention dakarand, but it is in the same class as havaged.



> TL;DR; it fetches random bytes over
> HTTPS to seed /dev/random. (They do pin the certificate in the client
> which is less crazy :).


they have plenty of central trust, so seeding over HTTPS perfectly
sane and a clear improvement for this situation.



> To be honest, I don't have a good way of fixing this issue. Feeding the
> urandom-seed with the date might be better than nothing but again I
> think that if a NTP correction occurs before seeding it, an attacker
> could end up knowing the seed if the NTP server or the link is
> malicious.


this is a valid concern and not resolved unless rngd with a good true
source is present. unfortunately, as you indicate, no good
alternatives...



> Is it crazy to think that Tails could provide a "seeding server" and use
> pollinate?


ideally a per image seed would be mixed before kernel calls init, like
device entropy mixing, along with a consensus runtime seed, before
calling init, but in a way that is all of mallory resistant, datagram
based, under fragmentation threshold, many transport distributed like
a Tor consensus over arbitrary IPv4, IPv6, DVB-T, clandestine
shortwave, and any other convenient local broadcast horizon. left as
exercise for reader. ;)


tls-tor-random to torproject, bridges, or onions as first step would
be quite convenient, in seriousness, for a better fix to poor entropy
situation at start. i have no idea how long just that part would be
to develop, what partitioning attacks it may expose its users to, nor
how much use it would receive in the wild, among other mysteries.


best regards,