>>>>> "Aaron" == Aaron Brown <abrown@eecs.harvard.edu> writes:

Aaron> On Sun, 17 Mar 1996, David Gilbert wrote:

	[my problem with ethernet, deleted]

Aaron> I've been having the *exact* same problem with the le interface
Aaron> on the Sparc/Sun4m port. I've spent countless hours trying to
Aaron> debug it since I thought it was a Sun4m specific
Aaron> problem...perhaps it is indicative of something more serious in
Aaron> shared code.

Aaron> Does everything behave fine until there is heavy traffic? If
Aaron> so, it sounds like the same problem. My current best theory is

	It's all relative.  I have two machines on my at-home network.
However, if I have more than one compile running on an nfs mounted
drive, then the problem shows up a lot more.  It also seems to appeear
a lot more if I'm doing something on the Sun4.

Aaron> that two interrupts come almost at once (i.e. a transmit and a
Aaron> receive interrupt); the driver notices both and services them
Aaron> at once, but locore delivers both interrupts anyway, so the
Aaron> second interrupt looks like a stray.

	This is one possibility, but...

Aaron> If this is the case, I have no idea why it would show up on
Aaron> both a 4/260 and an SS20...if anyone out there has ideas please
Aaron> feel free to share, as I would love to kick this bug!

	One problem that I had with the clock and the serial chips was
due to improper interupt locking... That is another possibility.
There is also the possibility that interupts are being generated for
some other reason (buffer full, etc.)?

Dave.

-- 
----------------------------------------------------------------------------
|David Gilbert, PCI, Richmond Hill, Ontario.  | Two things can only be     |
|Mail:      dgilbert@jaywon.pci.on.ca         |  equal if and only if they |
|http://www.pci.on.ca/~dgilbert               |   are precisely opposite.  |
---------------------------------------------------------GLO----------------