On Fri, 24 May 1996 13:00:37 +0300 (EET DST) 
 Jukka Marin <jmarin@muikku.jmp.fi> wrote:

 > > At 57600, we're talking about 1400us to respond before characters are
 > > lost.  Whats blocking for that long?  TTY is > then NET and BIO, right?
 > 
 > That's a long time, especially on a fast pentium CPU.  All I can tell is
 > that on my system, the wd interface doesn't seem to make the things
 > noticeably worse, but the ep (3com509) driver seemed to cause more data
 > loss.  The real problem may be somewhere else, though, because I got
 > FIFO overruns even when everything were supposed to be idle.

Note that TTY => NET => BIO assumes a "level" interrupt scheme, where the 
i386 uses a mask.  Because of slip/ppp/strip, TTY has to block NET.  The 
masks are calculated at run-time based on which devices ate attached to 
which IRQs.  Now, if your 3c509 shares the IRQ with your serial ports, 
NET will also block TTY, which could lead to lossage.

I think there are some IRQs which do have priority as far as interrupting 
the CPU ... but I'm not sure of the heirarchy.

----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