You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by gmu 2k6 <gm...@gmail.com> on 2006/07/25 16:46:04 UTC
building /dev/hwrng version of APR
what do you guys think about re-building APR to use use /dev/hwrng
instead of urandom or random (it's the default IIRC)?
it's on a HP ProLiant with Intel HW RNG in the chipset.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: building /dev/hwrng version of APR
Posted by gmu 2k6 <gm...@gmail.com>.
On 7/25/06, Andres Freund <an...@anarazel.de> wrote:
> Hi,
>
> On Tuesday 25 July 2006 20:28, gmu 2k6 wrote:
> > > > what do you guys think about re-building APR to use use /dev/hwrng
> > > > instead of urandom or random (it's the default IIRC)?
> > > I dont think that that makes much sense. You should use something like
> > > rngd from rng-utils that tests the random data and feeds it back into the
> > > kernel.
> > and it makes it available then via urandom or random? I guess it's random.
> Well, you can really seperate that. They are both fed from the same entropy
> pool. Which is fed from the rngd. The difference is just that urandom just
> gets seeded from the entropy pool and doesnt block on an empty pool.
well thanks for the help, I've installed rng-tools in Debian Etch and
it is meant to use /dev/hwrng which is existent. nice, I love it.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: building /dev/hwrng version of APR
Posted by Andres Freund <an...@anarazel.de>.
Hi,
On Tuesday 25 July 2006 20:28, gmu 2k6 wrote:
> > > what do you guys think about re-building APR to use use /dev/hwrng
> > > instead of urandom or random (it's the default IIRC)?
> > I dont think that that makes much sense. You should use something like
> > rngd from rng-utils that tests the random data and feeds it back into the
> > kernel.
> and it makes it available then via urandom or random? I guess it's random.
Well, you can really seperate that. They are both fed from the same entropy
pool. Which is fed from the rngd. The difference is just that urandom just
gets seeded from the entropy pool and doesnt block on an empty pool.
Greetings,
Andres
Re: building /dev/hwrng version of APR
Posted by gmu 2k6 <gm...@gmail.com>.
On 7/25/06, Andres Freund <an...@anarazel.de> wrote:
> Hi,
>
> On Tuesday 25 July 2006 18:46, gmu 2k6 wrote:
> > what do you guys think about re-building APR to use use /dev/hwrng
> > instead of urandom or random (it's the default IIRC)?
> I dont think that that makes much sense. You should use something like rngd
> from rng-utils that tests the random data and feeds it back into the kernel.
and it makes it available then via urandom or random? I guess it's random.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: building /dev/hwrng version of APR
Posted by Andres Freund <an...@anarazel.de>.
Hi,
On Tuesday 25 July 2006 18:46, gmu 2k6 wrote:
> what do you guys think about re-building APR to use use /dev/hwrng
> instead of urandom or random (it's the default IIRC)?
I dont think that that makes much sense. You should use something like rngd
from rng-utils that tests the random data and feeds it back into the kernel.
Greetings,
Andres