Help Phil debug your kernel crashes!

So you've tried one of my drivers and you've had the unfortunate experience of a kernel panic. You are somewhat distressed... but rather than just give up, you'd like to help me debug the driver you're trying out. To do that, you will need to send me a kernel stack trace.

Here's how to generate one.

First of all, this assumes that you had enough space in /var/crash to hold the kernel dump. It will have created two files,

/var/crash/unix.0
/var/crash/vmcore.0

[Adjust '0' as appropriate]

You then need to run the following commands as root:

cd /var/crash/`uname -n`
adb -k unix.0 vmcore.0

[It will eventually print out something like 'physmem f8b0']

$c

[control-d to get out of adb]

Send me the output of the $c, and I'll c what I can do! :-)
[but dont forget to also tell me which driver you are trying, which version of OS, what *specific* hardware, and sparc or x86]

More general crash-dump debugging

For extra details on more user-level triggered kernel dumps, try the following:

cd /var/crash/`uname -n`

crash -d vmcore.0 -n unix.0 
status
proc
kmastat

Written by: Philip Brown
Bolthole Top - Solaris Top - Driver Top - Search Bolthole.com