Subject: 3100 hangs - getting closer to a solution
To: Erik Bertelsen <erik@sockdev.uni-c.dk>
From: Michael L. Hitch <mhitch@lightning.oscs.montana.edu>
List: port-pmax
Date: 05/17/1997 15:25:42
On May 17, 10:35pm, Erik Bertelsen wrote:
> One further remark: I have this problem with a RZ56 as well. It sits in an
> expansion cabinet, and if I turn it off, the kernel will print a message
> on the console to tell that it went off line, but other than that the only
> thing to do is to press the reset or the power button.
>
> If I remember correctly, I could use this disk until about mid December,
> if this can help anyone to pin-point the problem.
>
> To some extent, I'll also be able to do some testing/debugging if that
> will help solving the problem.
Now that I've got a configuration that is easily hung, I've been
narrowing down where the hang is. I've narrowed it down to looping
in sii_DoIntr. It keeps looping back to the again: label. It appears
to happen when writing 64K bytes. I've added some code to detect the
loop, dump out the trace buffer, and reboot the system. I finally figured
out how to get the serial console to work and now have a copy of the console
output in a file. I should new be in a position to figure out what the
SII driver did up to where it hangs.
Something you may want to try is to disable disconnects: replace the
SCSI_DIS_REC_IDENTIFY with SCSI_IDENTIFY (around line 507 in sii.c). When
I tried that, it didn't appear to hang.
Michael
--
Michael L. Hitch mhitch@montana.edu
Computer Consultant
Information Technology Center
Montana State University Bozeman, MT USA