[blivet] My school project -- random data entropy in installation

Vratislav Podzimek vpodzime at redhat.com
Wed May 29 14:49:10 UTC 2013


On Tue, 2013-05-28 at 10:26 -0700, Brian C. Lane wrote:
> On Fri, May 24, 2013 at 07:01:41PM +0200, Vratislav Podzimek wrote:
> > These are patches I've come up with as part of my school project focused on
> > random data entropy in the installation process. The critical part that needs
> > high-quality random data is disk encryption. Patches 1/3 and 2/3 are, I believe,
> > useful in either case, patch 3/3 adds the entropy check and wait before the LUKS
> > format is created. Related Anaconda patches add a GUI dialog to inform user what
> > is going on and how they could help.
> 
> One thing I worry about is a kickstart being starved for entropy. Do you
> want to hang for however long it takes, do you want to disable waiting
> or provide a maximum timeout?
Good question. There might be some maximum timeout needed. However, I
think that the right threshold (which I'm still waiting for) should be
low enough not to hit any issues like that.

> Depending on the hardware being used you will have more (or less)
> entropy available. Last time I looked some of the network drivers
> contributed, but not all of them.
It's network drivers, HDD operations, keyboard, mouse, special
instructions HW, special instructions provided by the Ivy Bridge
processors etc.

-- 
Vratislav Podzimek

Anaconda Rider | Red Hat, Inc. | Brno - Czech Republic



More information about the anaconda-patches mailing list