On Sun, 25 May 1997 18:10:45 +1000 (EST) 
 Stephen.Ma+netbsd-tech-kern@jtec.com.au wrote:

 > I tend to agree with that last statement. However, I was thinking more
 > along the lines of moving most of the PPP code into user-land, and
 > using if_tun to pass up protocol packets (after LCP & NCP have been

"Ugh."

Think about the performance penalty ... I doubt that folks using
mutliple PPP connections (i.e. NetBSD PPP servers, etc.) would
appreciate paying the cost of moving the packets out to userland then
back into the kernel just to be forwarded.

What benefit does moving it all out into userland give you?  I seem to
recall that at one point the argument was "demand-dial", but seeing
as how we handle that even with a mostly in-kernel PPP, that argument
doesn't wash with me...

 > That would be nice. However, I'm a little worried about this approach
 > - the sync PPP if_sppp* code is a separate PPP stack to the async PPP
 > if_ppp* code from Paul Mackerras. I would rather have one PPP stack
 > that can handle both sync and async.

...as would I.  I would think that the sync PPP support could be added
to our PPP, and the changes be sent back to Paul.

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