Re: [Xen-users] pv_ops domU crashes on pv_ops dom0 (directly at boot)

[Jeremy Fitzhardinge <jeremy@xxxxxxxx>]

On 12/15/2009 04:49 AM, Markus Schuster wrote:
> On Wednesday 09 December 2009 22:35:44 Markus Schuster wrote:
>    
> > On Tuesday 08 December 2009 00:45:19 Jeremy Fitzhardinge wrote:
> >      
> > > On 12/06/09 14:08, Markus Schuster wrote:
> > > [..]
> > >
> > >        
> > > > What to do to find out the line in code that crashes?
> > > >          
> > > If you built the kernel, then:
> > > [..]
> > >
> > > You'll get best results if you enable CONFIG_DEBUG_INFO.
> > >        
> > Sorry for the delay, but I used a Debian binary kernel, so I had to grab
> > Debian sources and rebuild it in order to have reproduceable results. I've
> > enabled CONFIG_DEBUG_INFO from the start - it produced giant binaries :)
> >
> >
> > So I've done:
> > # gdb vmlinux
> > [..]
> > This GDB was configured as "x86_64-linux-gnu"...
> > (gdb) x/i 0xffffffff813861af
> > 0xffffffff813861af<cache_add_dev+46>:  mov    (%rdi),%r14d
> > (gdb)
> > (gdb) list *0xffffffff813861af
> > 0xffffffff813861af is in cache_add_dev
> > (/root/linux-2.6/linux-2.6-2.6.31/debian/build/source_amd64_none/arch/x86/k
> > ernel/cpu/intel_cacheinfo.c:902). 897     static
> >   DECLARE_BITMAP(cache_dev_map, NR_CPUS);
> > 898
> > 899     /* Add/Remove cache interface for CPU device */
> > 900     static int __cpuinit cache_add_dev(struct sys_device * sys_dev)
> > 901     {
> > 902             unsigned int cpu = sys_dev->id;
> > 903             unsigned long i, j;
> > 904             struct _index_kobject *this_object;
> > 905             int retval;
> > 906
> > (gdb)
> >      
> Hi Jeremy, sorry for beeing a bit cumbersome, but have you been able to get
> some findings from my debugger output? Do you need additional input?
>    
I don't understand what's happening here.  It looks like there's some 
kind of initialization order problem in sysfs, but I don't have any 
handle on the details here.
It would be interesting to compare the .config of the working kernel 
with the crashing one (using scripts/diffconfig in the kernel source 
tree) to see if there's some kernel option which may be related to the 
crash.
    J

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users