On Fri, 12 Jan 2007, Rui Paulo wrote:
> #if defined(I686_CPU) && defined(VIA_PADLOCK)

Done.

> /*
> + * VIA "Nehemiah" MSRs
> + */
> ...
> Maybe all these defines should be left padded?

Done, looks much better indeed.

> ---
>
> In via_padlock_attach() there's a memory leak if (vp_sc->sc_cid < 0).
>
> ---

Jikes, shouldn't have missed that, thanks!

> It would be great if you could replace bzero/bcopy by memset/memcpy (maybe 
> that should be in KNF).

Ok, done.

Thanks for the feedback!

-- Daniel