Ronald G Minnich <[EMAIL PROTECTED]> writes:
> On 6 Mar 2001, Eric W. Biederman wrote:
>
> > > makes sense. So my old theory about the kernel covering you were
> > > wrong. This sounds like a needed enhancement to page zero in kernels --
> > > making sure it's write-only.
> >
> > Hmm. I thought page zero was already read-only, but how that is
> > relevant currently escapes me.
>
>
> my typo. I meant to say 'page zeroing code'.
O.k. With the optimized code paths in the kernel it is the difference
between should never happen, and will never happen.
In the kernel what you want are the cache blocks for a page to become
dirty and zero as fast as possible. Which is a different case from
zeroing the SDRAM as fast as possible.
Having the ram set as a Write-Back instead of Uncached or
Write-Through I don't see how we could garantee that a
read-modify-write cycle never happens for cache blocks. So I the
kernel is probably o.k.
Eric
