: While we're on the topic of the 3c509, I can't get this blasted thing to 
: run for more than about 10 minutes before the "send packet" routine 
: wedges.  I am not even sure how to go about debugging this problem, and 
: I'm not very proficient with ddb (at least not yet) if that's the first 
: step I need to take anyway.  I've got
: 
: Micronics 486DX2-66 ISA/VLB main board
: 12MB RAM
: VLB IDE card of unknown origin
: Paradise Accelerator+ VLB SVGA

Well; fwiw, here is something that occured before I left for Boston Usenix:

I currently have several NetBSD machines running 3c509's.  The one I use
everyday, allday is a 486-66 EISA with an EISA BUslogic scsi controller
and an ISA S3-928 video card.

I managed to buy another box for other reasons at work so I bought a
486-66 ISA with 3 VLB slots, a VLB Buslogic scsi controller and a
VLB Mach32 video board. 

So; I transferred my 2 3c509's, and my 1.2GB disk into the new box and
powered it up.  One would think that everything should just continue
to work.  But I found that the 3c509's would occasionally hang and/or
have poor transfer rates.  Even if the rest of the system wasn't doing
much.  

Someone suggested I move the SCSI controller to a new VLB slot which I
did.  This seemed to increase dramatically the time between hangs (from
10 minutes to several hours) but did nothing for throughput.

Then it came time to go to Boston; so I swapped everything back and 
turned the VLB machine into a Windows/NT machine where it runs fine.

All this leads me to believe that VLB is actually a bug and is to be
avoided like the plague.  True that it is possible to make that
configuration work given that other operating systems do it, but the
fact that it exhibits itself on VLB systems,  indicates to me that there
is something inherently *wrong* with the design.  In Boston, I spoke
with Paul Vixie (who maintains the BSDI 3c5x9 driver) and he hinted that
the problems we were seeing were familiar to them.


hpeyerl@novatel.ca                           |  NovAtel Commnications Ltd.
hpeyerl@fsa.ca                               | <nothing I say matters anyway>
 "A sucking chest wound is nature's way of telling you to slow down."

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