Jonathan S. Shapiro
shap at eros-os.org
Mon Jan 16 13:27:06 EST 2006
On Mon, 2006-01-16 at 10:17 -0800, Charles Landau wrote:
> At 12:16 AM -0500 1/16/06, Jonathan S. Shapiro wrote:
> > On Sun, 2006-01-15 at 20:14 -0800, Charles Landau wrote:
> > > Then what problem are pinsets designed to solve?
> > Ensuring residency in the absence of a change in global policy.
> Aside from DMA, then, this is simply to ensure real-time response?
> Do you need to ensure not only that the page is resident (so no page
> fetch occurs), but that the address space map is set up (so no page
> fault occurs)? It could be in multiple address spaces. I came up with
> a solution to this at Tandem (see US Patent 6,279,095).
My current thinking is that fast page faults are probably okay.
One would think that after the factory patent you might have learned
your lesson. Shame on you.
More information about the coyotos-dev