Re: extra PG_* bits for page->flags

[David Howells <dhowells@warthog.cambridge.redhat.com>]

Valdis.Kletnieks@vt.edu wrote:
> > Is a new fs needed?  Is it not possible to use an existing filesystem of
> > the user's choice for local caching?
> 
> It's sort of like a loopback mount - you need a backing store which could be
> on an ext3 or ramfs or whatever, and you need the user-visible front end
> side of things, which manages the backing store and fetches blocks from
> AFS/NFS/etc.

Sort of, yes. Making it a filesystem is more a way of making it possible for
the kernel to associate a block device directly with the caching code. Rather
than adding an additional kernel interface beyond mount/swapon, you can just
use mount to add a block device to the cache.

Furthermore, it meand that I can provide virtual files along the lines of
/proc in the mountpoint that allow you to access status information, and maybe
also control the data cached.

David