| On Tue, Sep 25, 2001 at 02:41:04PM -0000, eeh@netbsd.org wrote:
|
| > This sounds like data corruption on the disk.  What kernel are you running?
| > What hardware is this?  I would suggest rebuilding the kernel with DMA turned
| > off to the disk, but obviously you can't do that.
|
| I've built and booted other kernels quite happily so I guess what
| you are suggesting is possible.
|
| This is a u10, running a -current from Sep 16.
|
| I tried another GENERIC build, which worked fine, but now..
|
| root device: wd1a
| dump device (default wd1b): 
| file system (default generic): 
| root on wd1a dumps on wd1b     
| root file system type: ffs
| warning: no /dev/console  
|
| then, I hit <enter> and..
|
| RED State Exception
|
| TL=0000.0000.0000.0005 TT=0000.0000.0000.0030
|    TPC=0000.0000.0100.8954 TnPC=0000.0000.0100.8958 TSTATE=0000.0000.5800.0504
| TL=0000.0000.0000.0004 TT=0000.0000.0000.0030
|    TPC=0000.0000.0100.8954 TnPC=0000.0000.0100.8958 TSTATE=0000.0000.5800.0504
| TL=0000.0000.0000.0003 TT=0000.0000.0000.0030
|    TPC=0000.0000.0100.8954 TnPC=0000.0000.0100.8958 TSTATE=0000.0000.5800.0504
| TL=0000.0000.0000.0002 TT=0000.0000.0000.0030
|    TPC=0000.0000.0100.8954 TnPC=0000.0000.0100.8958 TSTATE=0000.0000.5800.0504
| TL=0000.0000.0000.0001 TT=0000.0000.0000.0030
|    TPC=0000.0000.0127.8d7c TnPC=0000.0000.0127.8d80 TSTATE=0000.0099.8200.0604
|
| Anyone else seen similar on u5/10 with Sep 26 source?

Looks like your kernel stack may have gotten zorched.  What do you have at
kernel addresses 0100.8954 and 0127.8d7c?

| Oh, that reminds me - why must I specify the root device on every
| boot? Shouldn't it be passed from the prom to the kernel? I'm using:

That should not be necessary.  However, there are issues with identifying
some boot devices, most notably those on PCI SCSI controllers.

Eduardo