Subject: Kernel hang report: DDB info
To: None <port-sparc@NetBSD.ORG>
From: David Gilbert <dgilbert@jaywon.pci.on.ca>
List: port-sparc
Date: 03/02/1996 12:57:55
	I must say, than when I watch a machine boot, I always get a
feeling of awe --- at the process which has produced the beast
within.  However, of late, I have been watching this process a little
more often than I'd like, and this feeling has waned somewhat.

	In followup to the report that I'm experiencing regular hangs
when I try to reboot my machine, I have reproduced one of these with
my DDB kernel.

	I generally observe this after the machine has been up and
running for a fair amount of time with a reasonable load (news and
mail, uucp over the modem, etc).  In general, running for 18-24 hours
is enough to put the machine in the state that it will hang when I
type reboot.

	Now... It will print the syncing disks message --- and indeed,
I believe that it does.  It does not, however, set the fsclean bit ---
so that I have to wait for the fsck on bootup --- but when it hangs
like this, I have never had any errors or messages from the fsck that
would indicate that the fsck did not take place.

	This is more a nusiance than anything else, but there must be
some simple underlying problem.  The kernel in question is the result
of about a week old sup, with the recent si.c patch given to me by
Jason.

	Here's the 'trace' output from ddb, although I have deleted
the arguements since I was handwriting this down from the screen.  If
this turns out to be important, and someone tells me it is so, then I
can capture it next time this occurs.

	after writing this down, I typed continue.  I waited a few
more minutes (hit return a bunch along with CTRL-C) and then L1-A'd
again.  Trace output was identical:

_zstty () at _zstty+0x27c
_zstty () at _zstty+0x80
_sparc_interupt () at _sparc_interupt+0x104
_mi_switch () at _mi_switch+0x150
_tsleep () at _tsleep+1bc
_scheduler () at _scheduler+0x98
_main () at _main+0x500
_getidprom () at _getidprom+0x2a4

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