On Tue, 19 Dec 2000, John Summerfield wrote:
> [EMAIL PROTECTED] said:
> > Well, I don't see the benefit of using ring1 in stead of ring0. The
> > protection problem remains... 
> 
> Some things cannot be run in ring1; an attempt to so so generates a general 
> exception..

What things?  IIRC, ring0/1/2 are mostly identical except that you can
play with DPLs.  x86 has three supervisor levels and one user level.

> Then ring0 code can take a look and decide whether to emulate or kill, and if 
> emulation, then what kind of emulation (in some circumstances, 'device not 
> present" might be more appropriate than doing something a real device would).

This doesn't solve any of the performance issues with the current
system... it would just complicate it more.

-- Ramon


Reply via email to