mlockall(2 fixed

[Marius Strobl]

On Tue, Apr 27, 2004 at 06:54:37PM -0400, Brian Fundakowski Feldman wrote:
> Please try the following if you are having strange SIGBUS problems/leaked 
> wired memory and you are running programs that use mlockall(2).  There were 
> several distinct bugs that caused pages to never get unwired and completely 
> unrelated processes to act as if mlockall(2) was called when they had not 
> done so.
> 

This patch appears to fix the problems on the one box I could
reproduce it.