On Sun, 31 Dec 1995 15:48:49 -0800 
 Ted Lemon <mellon@fugue.com> wrote:

 > Then why isn't it called splmalloc(), or something like that?
 > splimp() implies to me something to do with networking.

*chuckle*

Actually, I was under the impression that "imp" in this case meant 
"important" not "internet message processor" :-)

 > Then if the lance driver is running for a long time at splimp(), it's
 > probably a bug that we need to fix.

On systems like the hp300, at least, you can't really avoid it; 
interrupts are heirarchial, and devices are configurable from ipl3 to 
ipl5 (6 is clock, 7 is NMI).  So, on the hp300, splbio, splnet, spltty, 
and splimp are all spl5; not raelly much choice in the matter.  (The 
clock doesn't use malloc()/free(), so splimp doesn't have to block it, 
which means that the heartbeat light on 400-series systems can blink 
during those operations :-)

Anyhow, it may be the case on the pmax, as well, that you just can't 
really avoid running at splimp in the LANCE code.  It's certainly what 
the hp300 does.

--------------------------------------------------------------------------
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