On Mar 2, 2006, at 5:42 AM, Matthias Drochner wrote:

>
> Now that the fault_type argument to uvm_fault() is
> a dummy everywhere and vm_fault_t has no meaning
> anymore outside uvm's internal use, we could clean
> up the interface a bit. Opinions?
> (and all calls from md trap.c would get the "0"
> stripped out)
>
> best regards
> Matthias

As I look at most of the changes, I just don't see
the point.  The changes just bloat the kernel a bit
without significantly simplifying things.  I would,
however, change vm_fault_t to an enum.