Re: 2.6.6-rc{1,2} bad VM/NFS interaction in case of dirty page writeback

[Christoph Hellwig]

On Wed, Apr 28, 2004 at 12:17:42PM -0400, Trond Myklebust wrote:
> On Wed, 2004-04-28 at 01:29, Christoph Hellwig wrote:
> > There's nothing speaking against probing for more dirty pages before and
> > after the one ->writepage wants to write out and send the big request
> > out.  XFS does this to avoid creating small extents when converting from
> > delayed allocated space to real extents.
> 
> You are referring to xfs_probe_unmapped_page()?

Yes, and the two other functions doing similar stuff.

> True: we could do
> that... In fact we could do it a lot more efficiently now that we have
> the pagevec_lookup_tag() interface.

Yes pagevec are much more efficient for that.  In fact I have a workarea
switching over XFS to use pagevec for that probing.

I'm not yet sure where I'm heading with revamping xfs_aops.c, but what
I'd love to see in the end is more or less xfs implementing only
writepages and some generic implement writepage as writepages wrapper.
-
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/