user-mode-linux-user

[uml-user] entropy hole in UML instances too

[Toralf F. <tor...@gm...>]

I think, that the discussed hole is related to UML too (or at least was
related in the past).
IIRC when I booted a freshly created UML image for the first time - it
hang in the boot phase for a looong time for the creation of
public/private keys of an applciation (either apache2 or courier, but I
do not remember).
While this behaviour itself was good (the application still waited b/c
/dev/random blocked I was explained) - it indicated too less entropy
during boot of UML too, isn't it ?

-- 
MfG/Sincerely
Toralf Förster
pgp finger print: 7B1A 07F4 EC82 0F90 D4C2 8936 872A E508 7DB6 9DA3

Re: [uml-user] entropy hole in UML instances too

[Paolo G. <p.g...@gm...>]

On Sat, Jul 7, 2012 at 11:58 PM, Toralf Förster <tor...@gm...> wrote:
> I think, that the discussed hole is related to UML too (or at least was
> related in the past).
> IIRC when I booted a freshly created UML image for the first time - it
> hang in the boot phase for a looong time for the creation of
> public/private keys of an applciation (either apache2 or courier, but I
> do not remember).
> While this behaviour itself was good (the application still waited b/c
> /dev/random blocked I was explained) - it indicated too less entropy
> during boot of UML too, isn't it ?

Yep - entropy comes from hardware interrupts, and UML doesn't have
enough such interrupts, as far as I remember; I think interrupts might
be individually marked as usable to produce entropy. I remember that
before I stopped development, a UML driver was introduced to allow UML
to feed on the host's entropy by reading the host's /dev/random.

Not sure about the current status of all this, though.

-- 
Paolo Giarrusso - Ph.D. Student
http://www.informatik.uni-marburg.de/~pgiarrusso/

Re: [uml-user] entropy hole in UML instances too

[richard -r. w. <ric...@gm...>]

On Sun, Jul 8, 2012 at 1:21 AM, Paolo Giarrusso <p.g...@gm...> wrote:

Nice to hear from you Paolo, are you planning to work again on UML? :)

> Yep - entropy comes from hardware interrupts, and UML doesn't have
> enough such interrupts, as far as I remember; I think interrupts might
> be individually marked as usable to produce entropy. I remember that
> before I stopped development, a UML driver was introduced to allow UML
> to feed on the host's entropy by reading the host's /dev/random.

It's CONFIG_UML_RANDOM.
Toralf, is it enabled in your config?

-- 
Thanks,
//richard

Re: [uml-user] entropy hole in UML instances too

[Toralf F. <tor...@gm...>]

On 07/08/2012 11:44 AM, richard -rw- weinberger wrote:
> On Sun, Jul 8, 2012 at 1:21 AM, Paolo Giarrusso <p.g...@gm...> wrote:
> 
> Nice to hear from you Paolo, are you planning to work again on UML? :)
> 
> 
> It's CONFIG_UML_RANDOM.
> Toralf, is it enabled in your config?
> 
Gah - it is not enabled ...

-- 
MfG/Sincerely
Toralf Förster
pgp finger print: 7B1A 07F4 EC82 0F90 D4C2 8936 872A E508 7DB6 9DA3