Subject: Re: tmpfs: Internal representation of data
To: Chuck Silvers <chuq@chuq.com>
From: Jason Thorpe <thorpej@shagadelic.org>
List: tech-kern
Date: 07/25/2005 10:14:55
On Jul 25, 2005, at 10:00 AM, Chuck Silvers wrote:

> being able to page out all of the per-file metadata seems like an
> important property of any file system, I would include that in the  
> design
> from the beginning.

Hm.

Perhaps tmpfs could provide a pool allocator that could be backed by  
pageable kernel memory.  So long as the tmpfs nodes are accessed only  
in a valid process context, all you need to do is simply access those  
data structures, and they could be paged in on demand.

(you could also explicitly wire and unwire the corresponding pages  
when the data structures are to be accessed.)

-- thorpej