Subject: sparcserver 470 problems
To: None <port-sparc@NetBSD.ORG>
From: Reini (Reinhold Huber) <Reinhold.Huber@Fachschaften.TU-Muenchen.DE>
List: port-sparc
Date: 02/03/1998 23:43:06
Hi,

some days ago I told about my problems with sparcserver 470.

Now I've compiled a kernel with both
SUN4_MMU3L and MMU3L enabled as an option (Not sure if I misspelled them, 
but I think you know what I mean :)

It seemed to me that the ss470 did not want to boot from something 
different than SCSI ID 3, so I jumpered the disk from 4 to 3, and so 
there was no "sd3" to mount on /usr/local, and so it tried to go to 
single user.
I pressed enter to get a 'sh', then there was a nice little kernel panic 
because of init dying, here are the console messages:

Feb 3 21:57:30 init: /bin/sh on /etc/rc terminated abnormally, going to 
single user mode
Enter pathname of shell or RETURN for sh:
Feb 3 21:57:37 init: can't exec P T for single user: No such file or 
directory
panic: init died (signal 4, exit 0)
syncing disks... done
Frame pointer is at 0xf44b1ca0
Call traceback:
  pc = 0xf012dd78  args = (0x0, 0x11000fe7, 0xf0170800, 0xf44b1dc0, 
0xf071a000, 0xf44b1d08) fp = 0xf44b1d08
  pc = 0xf0038474  args = (0x100, 0x0, 0x1, 0x0, 0xf44b1d40, 0xf44b1d3c, 
0xf44b1d70) fp = 0xf44b1d70
  pc = 0xf002a150  args = (0x100, 0x4, 0x0, 0x0, 0x20, 0xf44b1d70, 
0xf44b1dd8) fp = 0xf44b1dd8
  pc = 0xf003210c  args = (0xf0706400, 0x4, 0x80, 0xf000a33c, 0x114000c2, 
0xf44b1e58, 0xf44b1e40) fp = 0xf44b1e40
  pc = 0xf013b37c  args = (0xf0706400, 0x4, 0x38, 0xf44b0000, 0xf0009b78, 
0x20, 0xf44b1ea8) fp = 0xf44b1ea8
  pc = 0xf0008fc8  args = (0x6, 0x11100084, 0xa584, 0xf44b1fb0, 0x0, 0x3, 
0xf44b1f50) fp = 0xf44b1f50
  pc = 0xa578  args = (0x5, 0xeffffb48, 0x2, 0x0, 0x11100085, 0xf44b1fb0, 
0xeffffa80) fp = 0xeffffa80

dumping to dev 0x701, offset 241711
dump
Watchdog reset.
>

Questions:

1) What was the reason for the panic if this can be reconstructed from 
   the info above? (apart from that the fsck failed because of different 
   device names). I expected to get a single user shell to mount my 
   filesystems by hand, and there was a dump instead.
2) Has this anything to do with the difficulties that arise from the 
   special architecture of the ss470 (3-level MMU)?
3) What can I do to fix this? Pointers to documentation (preferably 
   Internet, I have no handbook for the machine) and other info should be 
   enough. If somebody likes to tell me the story, she/he saves a lot of 
   time for me :)
3) If this is ss470-related, or this is not, but I get other ss470 probs,
   how do I extract info from the dump?
   Here, I do not expect a tutorial, but a little help to do the RTFM.
4) What should I read to get started with NetBSD kernel programming? I 
   fear, this will be necessary, soon.

I'd really like to get NetBSD running on the ss470, and if I succeed, 
will surely send you the code that got it working.

If you know of anybody in the world who tries to do the same, please let 
me know. It is always a good feeling when one knows there are others with 
the same prob, and we could share both the effort and the success.

5) Another question: How related is this prob to the changes in the support 
   for different buses currently discussed on the list? Do these changes 
   affect the code that might have to be changed for the ss470 thing?

Greetings,
Reinhold Huber