Re: [rfc][patch] remove racy sync_page?

[Nick Piggin]

Nick Piggin wrote:
> Josef Sipek wrote:
>
> > On Mon, May 29, 2006 at 07:34:09PM +1000, Nick Piggin wrote:
> > ...
> >
> > > Can we get rid of the whole thing, confusing memory barriers and all?
> > > Nobody uses anything but the default sync_page
> >
> >
> >
> > I feel like I must say this: there are some file systems that live
> > outside the kernel (at least for now) that do _NOT_ use the default
> > sync_page. All the stackable file systems that are based on FiST [1],
> > such as Unionfs [2] and eCryptfs (currently in -mm) [3] (respective
> > authors CC'd). As an example, Unionfs must decide which lower file
> > system page to sync (since it may have several to chose from).
>
>
> OK, noted. Thanks. Luckily for them it looks like sync_page might
> stay for other reasons (eg. raid) ;)

Ah, no: that's the backing dev's unplug_fn. sync_page is always the
same. And it has definite bugs now that splice has been merged.

--
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/