actually, the correct way to fix this is to change the way we describe ranges
of address space so that we don't overflow the types involved.  then we won't
have any limitations on usable address space other than the hardware's.
that's a big change though, so I'd rather work around the problems for now
until we can fix it the right way.

-Chuck


On Wed, Jun 13, 2001 at 10:24:46PM -0700, Jason R Thorpe wrote:
> On Thu, Jun 14, 2001 at 08:12:57AM +0300, Chuck Silvers wrote:
> 
> The correct way to address this is to clip the MAX ranges in vmparam.h,
> not work around a *single* instance of this problem in UVM.
> 
>  > 
>  > Module Name:	syssrc
>  > Committed By:	chs
>  > Date:		Thu Jun 14 05:12:57 UTC 2001
>  > 
>  > Modified Files:
>  > 	syssrc/sys/uvm: uvm_fault.c
>  > 
>  > Log Message:
>  > work around an overflow problem in uvm_fault_wire().
>  > from Eduardo Horvath and Simon Burge.
>  > 
>  > 
>  > To generate a diff of this commit:
>  > cvs rdiff -r1.64 -r1.65 syssrc/sys/uvm/uvm_fault.c
>  > 
>  > Please note that diffs are not public domain; they are subject to the
>  > copyright notices on the relevant files.
>  > 
> 
> -- 
>         -- Jason R. Thorpe <thorpej@zembu.com>