Messages in this thread | | | Date | Wed, 12 Dec 2012 20:10:16 +1100 | From | NeilBrown <> | Subject | Re: [RFC] AES instead of SHA1 for /dev/urandom |
| |
On Wed, 12 Dec 2012 09:58:16 +0100 Ondřej Bílka <neleai@seznam.cz> wrote:
> On Wed, Dec 12, 2012 at 01:08:26PM +1100, NeilBrown wrote: > > On Wed, 12 Dec 2012 03:03:54 +0100 Ondřej Bílka <neleai@seznam.cz> wrote: > > > > > I consider to speed-up /dev/urandom on recent intel processors by > > > using hardware aes. Same for accelerated aes crypto. > > > > > > Would you accept a patch if I wrote it? > > > > Have you read https://lwn.net/Articles/525459 ? > > > Yes > > In particular the paragraph containing: > > > > A member of the audience asked why the kernel couldn't just do away with > > the existing system and use the HWRNG directly. > > > > Does that answer your question in any way? > > > No as AES is a HWRNG about as much as horse is type of automobile.
Yes, of course. Thanks.
NeilBrown [unhandled content-type:application/pgp-signature] | |