[CUWiN-Dev] Re: Fail to boot after upgrade

[Lu-chuan Kung]

David Young <dyoung <at> pobox.com> writes:

> 
> On Tue, Oct 31, 2006 at 03:58:16PM -0600, Lu-chuan Kung wrote:
> > Hi,
> > 
> > I've encountered a weird problem after upgrading node image on our 
> > net4826 boards. I'm sure the image file is good because there are other 
> > nodes who are running on the same image. The boot loader have a 
> > "Input/output error" when loading the kernel image (/netbsd) on 
> > /dev/wd0e. I checked the kernel image file by diff and it looks just 
> > fine. I've noticed that all the problmatic boards have the kernel image 
> > on /dev/wd0e instead of /dev/wd0a. Can anyone tell me what might cause 
> > the problem? The following are the booting messages I got after upgrade:
> 
> Luke,
> 
> I have a hunch that the geometry of the original image and the geometry
> of the CF card are mismatched.
> 
> If you hold down Spacebar or some other "innocuous" key during boot,
> you can get the bootloader to drop to the boot menu.  Then you can type
> 'boot hd0a:netbsd' to boot the system on the old partition.  Log in as
> root (or su to root) and run 'disklabel wd0'.  Send us the output.
> 
> Also, send us the output of the Soekris BIOS during boot-up.
> 
> Dave
> 


Hi David,

Here are the output of "disklabel wd0":
-------------------------------------------------------------------------
# /dev/rwd0d:
type: ESDI
disk: net4526
label: 
flags:
bytes/sector: 512
sectors/track: 32
tracks/cylinder: 4
sectors/cylinder: 128
cylinders: 977
total sectors: 125056
rpm: 3600
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0           # microseconds
track-to-track seek: 0  # microseconds
drivedata: 0 

5 partitions:
#        size    offset     fstype [fsize bsize cpg/sgs]
 a:     62512        32     4.2BSD      0     0     0  # (Cyl.      0*-    
488*)
 c:     62512        32     unused      0     0        # (Cyl.      0*-    
488*)
 d:    125056         0     unused      0     0        # (Cyl.      0 -    976)
 e:     62512     62544     4.2BSD   1024  8192  7816  # (Cyl.    488*-    976)
---------------------------------------------------------------------------


The output of "fdisk wd0":
---------------------------------------------------------------------------
# fdisk wd0
Disk: /dev/rwd0d
NetBSD disklabel disk geometry:
cylinders: 977, heads: 4, sectors/track: 32 (128 sectors/cylinder)
total sectors: 125056

BIOS disk geometry:
cylinders: 977, heads: 4, sectors/track: 32 (128 sectors/cylinder)
total sectors: -9035909694451351552

Partition table:
0: NetBSD (sysid 169)
    start 32, size 62512 (31 MB, Cyls 0-488/2/17)
1: NetBSD (sysid 169)
    start 62544, size 62512 (31 MB, Cyls 488/2/17-977), Active
2: <UNUSED>
3: <UNUSED>
Bootselector disabled.
----------------------------------------------------------------------------


BIOS Booting Screen
----------------------------------------------------------------------------
comBIOS ver. 1.28  20050529  Copyright (C) 2000-2005 Soekris Engineering.

net4801

0064 Mbyte Memory                        CPU Geode 233 Mhz

Pri Mas     64MB  ATA Flash Disk         LBA 977-4-32  62 Mbyte

Slot   Vend Dev  ClassRev Cmd  Stat CL LT HT  Base1    Base2   Int
-------------------------------------------------------------------
0:00:0 1078 0001 06000000 0107 0280 00 00 00 00000000 00000000
0:06:0 100B 0020 02000000 0107 0290 00 3F 00 0000E101 A0000000 10
0:14:0 168C 0013 02000001 0116 0290 08 3C 00 A0010000 00000000 11
0:15:0 168C 0013 02000001 0116 0290 08 3C 00 A0020000 00000000 05
0:18:2 100B 0502 01018001 0005 0280 00 00 00 00000000 00000000
0:19:0 0E11 A0F8 0C031008 0117 0280 08 38 00 A0030000 00000000 09

 1 Seconds to automatic boot.   Press Ctrl-P for entering Monitor.
----------------------------------------------------------------------------

I noticed that in the output of disklabel. The number of total sectors of BIOS 
disk geometry is negative. However, the individual parameters (Cyl,Head,S/Cyl) 
seem to be correct and thus I couldn't fix it in disklabel.

And by the way, why do we need disk geometry when using "/sbin/upgrade"? I 
didn't notice anything related to disk geometry in the source of the script.

Luke