Subject: Machine check/ fatal kernel trap on 1.5.1 Multia
To: NetBSD/Alpha Mail List \(E-mail\) <port-alpha@netbsd.org>
From: David Woyciesjes <DAW@yalepress3.unipress.yale.edu>
List: port-alpha
Date: 08/02/2001 16:38:08
Well, here's the story...
	I loaded up 1.5.1 on my 166MHz Multia (192 MB RAM), and built a
kernel (actaully 2) for it, with patches from Juergen Weiss
<weiss@Uni-Mainz.DE>...
http://mail-index.netbsd.org/port-alpha/2001/07/20/0017.html
... They (the kernels) work fine, using the console.

# To keep this mesage short, full output details and kernel configs are
here...
http://pages.cthome.net/squeege/terminal/

Part 1:
	When I hooked up a VT420 (9600/8/none/1/VT220 term type) to the
serial port, I get a fatal kernel trap...
.....
pcmcia0 at pcic0 controller 0 socket 0
pcmcia1 at pcic0 controller 0 socket 1

fatal kernel trap:

     trap entry = 0x2 (memory management fault)
     a0         = 0x0
     a1         = 0x1
     a2         = 0x0
     pc         = 0xfffffc00003669b0
     ra         = 0xfffffc000036699c
     curproc    = 0xfffffc00006acdd8
         pid = 0, comm=swapper

panic: trap
Stopped in swapper at  cpu_Debugger+0x4:     ret     zero,(ra)
db>

.....
	So I hooked it up to my Win98 box, and used Hyperterminal
(9600/8/none/1/autodetect term type). It booted fine, but when I issued
reboot, I got...
.....
CPU 0 booting


Unexpected Machine Check through vector 00000067

IPRs:
EXC_ADD:00000000001303A0  ICCSR:   0000000000000000  HIER:
000000001FFFDC70
.....
	Plug the keyboard back into the Multia for console, and it comes up
fine. Now, here's how the two bombs relate. When I issue reboot at the db
prompt on the VT420, the same machine check scrolls by.
	Now, I know a 660 vector usually means memory ECC error, or other
hardware error, but why does it work fine on console?

Part 2:

I get this, when the Multia boots up with the VT420...
.....
pci0: i/o space, memory space enabled, rd/line, rd/mult, wr/inv ok
unknown vendor 0x2041 product 0x1f01 (class 0x1f, subclass 0x01, interface
0x21d
<<<4 same lines clipped>>>
unknown vendor 0x2041 product 0x1f01 (class 0x1f, subclass 0x01, interface
0x21d
siop0 at pci0 dev 6 function 0: Symbios Logic 53c810 (fast scsi)
.....
and this with Hyperterminal...
.....
pci0: i/o space, memory space enabled, rd/line, rd/mult, wr/inv ok
unknown vendor 0x2041 product 0x1f01 (class 0x1f, subclass 0x01, interface
0x20, revision 0x41) at pci0 dev 0 function 0 not configured
<<<4 similar lines clipped>>>
unknown vendor 0x2041 product 0x1f01 (class 0x1f, subclass 0x01, interface
0x20, revision 0x41) at pci0 dev 5 function 0 not configured
siop0 at pci0 dev 6 function 0: Symbios Logic 53c810 (fast scsi)
.....
yet, here's the same (booting) area on console
.....
pci0 at lca0 bus 0
pci0: i/o space, memory space enabled, rd/line, rd/mult, wr/inv ok
siop0 at pci0 dev 6 function 0: Symbios Logic 53c810 (fast scsi)
siop0: interrupting at isa irq 11
.....
	The only difference I see betweeen the VT420 and Hyperterm, is the
interface number, 0x21d vs. 0x20. It seems like it's a configuration issue
with the two terminals, but I'm haven't the foggiest idea where...

***Full output details and kernel configs are here...
http://pages.cthome.net/squeege/terminal/

---   David A Woyciesjes
---   C & IS Support Specialist
---   Yale University Press
---   mailto:david.woyciesjes@yale.edu
---   (203) 432-0953
---   ICQ # - 905818