Re: Problems with mmap consistency

[Olaf Kirch <okir@suse.de>]

On Fri, Feb 24, 2006 at 05:01:42AM +0100, Andrea Arcangeli wrote:
> > We need to add a VM helper that will invalidate the page cache and flush
> > dirty pages atomically (such a thing does not yet exist).
> 
> Such thing will hardly exist if page faults can mark pages dirty at any
> time mostly lockless. Things may change for the better if we start
> taking the page lock in do_no_page but I doubt we'll serialize
> atomically against the whole disk-bound fdatawriteandwait.

One thing I've been thinking about, which would be rather heavy-handed
though, would be to have an rwsem in the mapping; do_no_page would take
a readlock, while nfs would take the write lock when invalidating.
Bu that's really ugly because we'd add this lock for NFS only.

> I think it's enough to make invalidate_inode_pages2 non destructive.

That would be worth trying. I'll give your patch a try.

Olaf
-- 
Olaf Kirch   |  --- o --- Nous sommes du soleil we love when we play
okir@suse.de |    / | \   sol.dhoop.naytheet.ah kin.ir.samse.qurax


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs