On Sat, Jan 31, 2004 at 08:36:54AM +0900, enami tsugutomo wrote:
> > as it should, but at least this way it will see the state of the
> > free count and free target _before_ the scan step does its thing.
> 
> Now the scan step will see the number after some pages are back, won't
> it?

Yes, it will, and the overall memory balance between the buffer cache
and the page cache appears to be far more stable that way.  With the
buf_drain(0) at the end of the pagedaemon, no buf_canrelease() criterion
using the free target turned out to work well in practice because the
scan almost always meets the target, or gets very close; with the
buf_drain(0) at the head of the pagedaemon, the scan still runs most of
the time, though I must confess that I am not entirely sure _why_ it
works better this way.

-- 
 Thor Lancelot Simon	                                      tls@rek.tjls.com
   But as he knew no bad language, he had called him all the names of common
 objects that he could think of, and had screamed: "You lamp!  You towel!  You
 plate!" and so on.              --Sigmund Freud