Panic on FreeBSD 6.0BETA1

[Panagiotis Christias]

On 8/9/05, Kris Kennaway <kris at obsecurity.org> wrote:
> On Wed, Aug 03, 2005 at 08:47:43AM -0400, Derrick Edwards wrote:
> > ???Hi all,
> > I decided to try and help with testing 6-BETA1, after updating sources and
> > recompiling i get the following during boot up. I am tried booting without
> > hyperthreading enabled in the bios and I still get the same panic.
> >
> > Fatal trap 12: page fault while in kernel mode
> > cpuid = 0; apic id =00
> > fault virtual address = 0x480008
> > fault code = supervisor read, page not present
> > instruction pointer = 0x20:0xc06923cc
> > stack pointer = 0x28:0xc10208ec
> >
> > Code segment = base 0x0, limit 0xfffff, type 0x1b
> > ? ? ? ? ? ? ?= DPL 0, pres 1, def32 1, gran 1
> >
> > processor eflags = interrupt enabled, resyne, IOPL = 0
> > current process = 0 (swapper)
> > [thread pid 0tid0]
> > Stopped at strlen+0x8: cmpb $0,0(%edx)
> >
> > I left all the debugging features in the current however, I am not sure
> > exactly how to trace this problem. If someone could point me to any doc that
> > would allow to provide more information that would ?be great. I updated my
> > pen 400MHZ using the same procedures and all went well. Please help
> 
> See the chapter on kernel debugging in the developers handbook.
> 
> Kris
> 
> P.S. I need to make a macro to type the answer to this FAQ.

IMHO I think that the average FreeBSD user/fan could use some help
when he/she deals with kernel panics/crashes/dumps/kgdb/kernel.debug
etc.

First step, the debug kernel and modules could be included in the base
system. What is it, 50 to 60MB disk space? That should be no(?)
problem. Being there and correctly built by "The Source" would be the
first step for the average user (not the "I wrote the kernel"
developer) to use them and provide correct, complete and thus useful
feedback. The one that you developers ask for.

Second step, the procedure of extracting the useful data from the
crash dumps using the "wat doez diz program do?" kgdb program is not
efficient to be left in hands of the average user. He/she is probably
a great guy, afterall he/she is using FreeBSD and certainly willing to
contribute but dealing with kgdb is too much to ask for. Except if
he/she has the guidance of kernel developer through this mailing list
or others. It is not an efficient way to investigate and report
serious problems.

I am wondering if possible to write some kgdb-wrapper scripts which
would do the dirty job of analyzing the crash dump using kgdb and the
proper command (backtrace, list xx, print yy, up nn, down mm, voodoo
etc.). This procedure can be as detailed as needed or wanted.  The
results could be sent automatically (send-pr) to the FreeBSD Team.

That way, every user would be able to easily send proper reports
without mistakes or missing data and benefit the FreeBSD Project.

Any thoughts?
Panagiotis