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

[Nathan Scott <nathans@sgi.com>]

On Wed, Apr 28, 2004 at 12:07:15PM -0700, Andrew Morton wrote:
> Christoph Hellwig <hch@infradead.org> wrote:
> >
> > 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.
> ...
> Any code which implements writearound-inside-writepage should be targetted
> at a generic implementation, not an fs-specific one if poss.  We could go

Another thing I think XFS could make good use of on the generic
buffered IO path would be mpage_readpages and mpage_writepages
variants that can use a get_blocks_t (like the direct IO path)
instead of the single-block-at-a-time get_block_t interface thats
used now.  The existence of such beasts would probably impact the
direction xfs_aops.c takes a fair bit I'd guess Christoph?

cheers.

-- 
Nathan