> > one potential problem here:  we shouldn't ever put the pagedaemon to
> > sleep waiting for memory which is created by the pagedaemon.  this
> > wasn't a problem before since the buffer cache memory was made available
> > for reuse in a different buffer by brelse(), ie. usually when i/o completed.
> 
> Under which circumstances could this actually happen?  The page daemon
> doesn't rely on the buffer cache directly. I guess there's a remote
> possibility to trigger a buffer cache allocation when pageing out pages
> backed by a file that's mapped in shared mode.

buffer caches are needed for metadatas.

YAMAMOTO Takashi