On Sun, Sep 01, 2002 at 09:06:58PM -0700, Chuck Silvers wrote:

 > your diagnosis is right, however for the fix I think it would be better
 > to handle the O->K case in uvm_fault() rather than in uvm_pagerealloc().
 > as it stands, the latter doesn't know anything about loaning, and
 > it's probably better to keep it that way.

Ok, tomorrow I'll adjust the patch and check it in.

Thanks for the sanity check!

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>