Re: [PATCH] pagecache lock ordering

[Andrew Morton]

Benjamin LaHaise wrote:
> 
> On Thu, Jan 10, 2002 at 12:39:11PM -0800, Andrew Morton wrote:
> > I'm struggling to see a use for generic_buffer_fdatasync().  Maybe
> > for a filesystem which doesn't implement ->writepage()?  Dunno.
> 
> I seem to be using it in aio.  Well, at least code based on it which
> seems to work for most filesystems for O_DATASYNC...
> 

You seem to be using writeout_one_page().  What I was
thinking of was:

- Kill generic_buffer_fdatasync().
- Move writeout_one_page() into fs/buffer.c
- Move waitfor_one_page() into fs/buffer.c.  This is just
  for completeness; I expect this function will have no
  callers soon.  __iodesc_sync_wait_page() could use it though.

OK by you?

-