On Wed, 10 Jan 1996 06:04:07 -0600 
 Michael Graff <explorer@flame.org> wrote:

 > Jan 10 04:25:43 packrat /netbsd: ed0: device timeout
 > 
 > packrat has a different card -- an NE2000 compatible -- so the only common
 > element other than the arch and the isa bus is the driver and its interaction
 > with gcc 2.7.2 (not 7.2.7 ;)
 > 
 > Has anyone else noticed this?  I compiled both kernels with -O, not -O6,
 > so that shouldn't be the famous -fstrength-reduce bug, right?

It could very well be that you're seeing interrupt lossage resulting from 
the way gcc 2.7.2 generates code; there was a discussion about it here 
recently.  I'm pretty sure Charles fixed it.

I'm goingto be updating my i486 with an SMC (ed) at work today (now that 
the disk stuff is checked in) and so I'll be able to confirm whether or 
not this is fixed or still a problem...  I'll be using Jan 9 source.

ciao.

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