Re: Key operations in the Kernel Crypto framework

On Fri, Jan 10, 2003 at 01:30:37AM -0700, Theo de Raadt wrote:
> > There has not (to my knowledge) been any reason for the kernel to do these
> > kinds of operations, as opposed to doing symmetric crypto.
>
> You do not want kernel code do to these operations. The kernel is

Well, it's either that or farm it out to a user-space daemon, which then results in a bootstrapping problem. There's also the problem of telling it what image to validate, without having races.

> non-preemptive. If you don't know what that means, just take my word

Because of course the average user doesn't have enough sense to operate google in an appropriate manner. Pfft.

Seriously, The whole non-pre-emption part of things is part of the challenge of what I'm doing. The challenge is finding the appropriate points at which to schedule.

-- 
Ceri Storey