On Wed, 11 Sep 96 17:55:58 EDT 
 "Gordon W. Ross" <gwr@mc.com> wrote:

 > (2) Come up with a way for driver interrupt routines to wait for
 >     the required DVMA space before starting the next transfer.
 > 
 > I think I can do (1) fairly easily, and that will allow up to 3
 > controllers or so.  Doing (2) would be better, but would really
 > need kernel threads to do nicely, so the driver could just use a
 > "work" thread to do transfers, so the "work" thread can block on
 > resource shortages (like DVMA space).  That's a long way off...

Maybe a good short term solution is to set up a timeout to restart the 
work queue in the event it gets stalled due to DVMA space shortage?

 -- save the ancient forests - http://www.bayarea.net/~thorpej/forest/ -- 
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939