Re: [PROPOSAL/PATCH] Fortuna PRNG in /dev/random

[Lee Revell]

On Fri, 2004-09-24 at 14:43, James Morris wrote:
> On Fri, 24 Sep 2004, Theodore Ts'o wrote:
> 
> > have *any* encryption algorithms in the kernel at all.  As to whether
> > or not cryptoapi needs to be mandatory in the kernel, the question is
> > aside from /dev/random, do most people need to have crypto in the
> > kernel?  If they're not using ipsec, or crypto loop devices, etc.,
> > they might not want to have the crypto api in their kernel
> > unconditionally.
> 
> As far as I know embedded folk do not want the crypto API to be mandatory,
> although I think Matt Mackall wanted to try and make something work
> (perhaps a subset just for /dev/random use).

/dev/random used to be a source of high latencies, but Ingo's patches 
fix this.  There was not a lot of CPU overhead but the latency was was a
problem for serious audio use.  But, audio is a unique set of
requirements, it's somewhere between desktop and embedded and hard-RT.

This could certainly be a problem for the embedded folks due to space or
CPU concerns, but the latency problem seems to be solved.

Lee  

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/