Re: PATCH: rev_pte_1 -- please test

["Stephen C. Tweedie" <sct@dcs.ed.ac.uk>]

Hi Ben,

A first quick comment on the new vma stuff:

Hmm, you've overloaded the vma/vm_offset stuff with the inode queues!
That's OK in principle --- the swap cache stuff currently maintains both
the per-inode page lists AND the hash lists, but only the hash lists are
strictly necessary so I guess we can live with that.  

There's also the guts of a page queue implementation --- have you got
any firmer plans for that?  I'm still uncertain about the benefits of
having these queues, except for the obvious use of the dumpable list.

Cheers,
 Stephen.