Subject: Re: kern/21107: genfs_putpages modifies page tables without proper locking
To: Brian Buhrow <>
From: Manuel Bouyer <>
List: netbsd-bugs
Date: 04/16/2003 09:58:32
On Tue, Apr 15, 2003 at 11:08:34PM -0700, Brian Buhrow wrote:
> 	Hello Manuel.  No, I'm not using any of those options.  I'm not sure
> why the patch works either.  Could it be that the compiler changes the
> alignment of something, even if the definition is null?  

No it should not. It's a #define, so the compiler itself won't see anything.
You can probably confirm this with cc -s

Maybe something else changed in your kernel ?

> 	It seems like a good idea to incorporate the fix, though so that when
> we do use lockdebug or multiprocessor, things work. :)

If I understood properly, this change should not be needed even for 
multiprocessor (pmap_clear_modify() doesn't need to have the page queue

Manuel Bouyer, LIP6, Universite Paris VI. 
     NetBSD: 24 ans d'experience feront toujours la difference