Re: port-xen/47720: Kernel Panic on dom0 shutdown or reboot when using RAIDframe

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Fri, Apr 05, 2013 at 11:15:01PM +0000, kimtwain0%gmail.com@localhost wrote:
> [...]
> >How-To-Repeat:
> 1. Install NetBSD 6.0.1 on wd0
> 2. Create a RAIDframe array (raid1) using wd1 and wd2
> 3. Install xentools41, xenkernel41 and extract netbsd-XEN3_DOM0.gz (version 
> 6.0.1) as /netbsd.dom0
> 4. Add 
> "xenbackendd=YES
> xend=YES
> xendomains=YES
> xenwatchdog=YES" to /etc/rc.conf
> 5. Create a new entry in /boot.cfg menu=Xen:load /netbsd.dom0 
> console=pc;multiboot /usr/pkg/xen41-kernel/xen.gz dom0_mem=256M
> 6. Reboot into xen
> 7. Issue a shutdown or a reboot

All my Xen servers have multiple raidframe on wd?, and I've never
seen this, so there must be something else.

On Sat, Apr 06, 2013 at 01:35:02AM +0000, Kim Twain wrote:
>  kernel: page fault trap, code=0
>  stopped in pid 34.1 (halt) at netbsd:bus_space_read_4+0x8: movl 0(%rdx),%eax
>  bus_space_read_4() at netbsd:bus_space_read_4+0x8
>  Xresume_xenev6() t netbsd:Xresume_xenev6+0x47
>  --- interrupt ---

Xresume_xenev6() doesn't call bus_space_read_4(), there is other
functions in between, and this is the information missing.
Can you try to rebuild the kernel with -O0, and maybe some -fno-...
options, like -fno-optimize-sibling-calls


-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--