Subject: weird messages keep popping up
To: None <port-i386@NetBSD.ORG>
From: Phil Knaack <flipk@idea.exnet.iastate.edu>
List: port-i386
Date: 09/12/1996 21:24:43
I am having some really strange crashes and messages on my 486 DX2/50,
wondering if someone can interpret these for me:

	Data modified on freelist: word 3 of object 0xf87b2ec0
		size 56 previous type VM mapent (0xdead9eef != 0xdeadbeef)

First, I'd like to point out that the address "0xdeadbeef" is REALLY odd.
I'm subscribed to a humor/news mailing list by that same name. Why is this
showing up?

After that appears, the machine continues to run for a few more seconds,
then, 

	panic: cluster_callback: too little memory

and at that point, the result of a 'trace' is (addresses removed):

	biodone
	scsi_error
	scsi_err1
	scsi_done
	aha_done
	aha_finish_ccbs
	ahaintr
	Xrecurse11
	   ----
	idle
	bpendtsleep
	biowait
	cluster_read
	ffs_read
	vn_read
	sys_read
	syscall
	  ---- syscall #3 ----

I assume what this means is that something failed while something was being
read from the disk.  This is somewhat coincidental to the occurance of me
reconnecting my ZIP disk to my aha1542 some days ago (and I had a lot of
problems getting it to work for a time). Do I have bad cabling? Is my
controller fried? Or is something hosed with my setup, or is there a bug
somewhere that's just now being triggered?

My setup is 1.2_BETA, circa June 30. Its been running perfectly stable
since this last build.

After rebooting the machine (takes two 'call boot's, it fails the first 
time while trying to sync()), fsck reports errors such as

	LINK COUNT FILE I=xxx -255 SHOULD BE 1

a couple of times, its random which partition, and it should be noted that
this has occured on both my SCSI disk (Root partition) and on an IDE disk.

Also I've had errors pop up like

	UNDEFINED OPTIMIZATION IN SUPERBLOCK
	SET TO DEFAULT?

and even though I may answer with "y", an immediate repeat of the "fsck"
command brings the same error again. Miraculously, after the last crash,
this particular error finally went away.

What's going on? Any ideas?

Cheers,
Phil
--
Phillip F Knaack
Systems Administrator, Information Development for Extension Audiences (IDEA)
Iowa State University Extension