Re: [patch 4/8] mm: allow not updating BDI stats in end_page_writeback()

[Andrew Morton <akpm@linux-foundation.org>]

On Tue, 18 Mar 2008 09:11:49 +0100 Miklos Szeredi <miklos@szeredi.hu> wrote:

> > 
> > Reason: bdi_cap_writeback_dirty() is kinda weirdly intrepreted to mean
> > various different things in different places and we really should separate
> > its multiple interpretations into separate flags.
> > 
> > Note that this becomes a standalone VFS cleanup patch, and the fuse code
> > can then just use it later on.  
> 
> Hmm, I can see two slightly different meanings of bdi_cap_writeback_dirty():
> 
>  1) need to call ->writepage (sync_page_range(), ...)
>  2) need to update BDI stats  (test_clear_page_writeback(), ...)
> 
> If these two were different flags, then fuse could set the
> NEED_WRITEPAGE flag, but clear the NEED_UPDATE_BDI_STATS flag, and do
> it manually.
> 
> Does that sound workable?

Yup, thanks.