This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


[Xen-users] Debian DomU with physdev access init kernel crash ....

To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-users] Debian DomU with physdev access init kernel crash ....
From: Himanshu Raj <rhim@xxxxxxxxxxxxx>
Date: Sun, 18 Dec 2005 12:08:28 -0500
Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Sun, 18 Dec 2005 17:11:04 +0000
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Mutt/1.5.11
Hi Folks,

I saw some references to the same problem in mailing list but no solution. 
I am running a DomU with PHYSDEV access. With fedora core 4, I just changed
the inittab file to not start getty's on tty1-6, and it worked (only ttyS0
is enabled). However, same trick didn't do it for debian. I can boot into a
shell if I put init=/bin/bash on kernel command line. However, starting init
with any runlevel bombs. Same fixes as FC4 in the inittab file didn't do it.
So I removed the whole init.d/rcS with a much simpler rcS similar to that of
ttylinux-xen. However, that doesn't help at all. My suspicion is that the
init process itself (before reading any of the setup stuff) is trying to open
a virtual console. 

Can anyone point me out to a solution to this problem? Or anyway to debug this

Thanks and best regards,
P.S. - This is the kernel crash dump. The call to visual_init is giving it 

Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
 *pde = ma 00000000 pa 55555000
 Oops: 0000 [#1]
 Modules linked in:
 CPU:    0
 EIP:    0061:[<c02052b3>]    Not tainted VLI
 EFLAGS: 00010246   ( 
 EIP is at visual_init+0x53/0x150
 eax: 00000000   ebx: c1114400   ecx: 00000000   edx: 00000000
 esi: 00000000   edi: 00000000   ebp: c003be84   esp: c003be68
 ds: 007b   es: 007b   ss: 0069
 Process init (pid: 1, threadinfo=c003a000 task=c0367a20)
 Stack: c1170000 c03dcc00 c03e35f8 c003bec4 c1114400 00000000 00000001 c003beac 
        c020543f c1114400 00000000 00000001 00000000 00000000 00000000 c1170000 
        00000000 c003bec4 c02083d8 00000000 ffffffed c1138b80 00000001 c003bef4 
 Call Trace:
        [<c0109faf>] show_stack+0x7f/0xa0
        [<c010a14f>] show_registers+0x15f/0x1e0
        [<c010a384>] die+0xf4/0x180
        [<c01155e4>] do_page_fault+0x4b4/0x761
        [<c0109c37>] error_code+0x2b/0x30
        [<c020543f>] vc_allocate+0x8f/0x150
        [<c02083d8>] con_open+0x48/0xb0
        [<c01f8b60>] tty_open+0x1b0/0x330
        [<c016d3f3>] chrdev_open+0xb3/0x1a0
        [<c016271b>] dentry_open+0x19b/0x210
        [<c016257a>] filp_open+0x5a/0x60
        [<c0162996>] sys_open+0x46/0x90
        [<c0109a81>] syscall_call+0x7/0xb
        Code: 05 01 d0 ff 88 e0 00 00 00 83 3a 02 0f 84 c6 00 00 00 8b 15 6c 9b 
35 c0 89 53 28 8b 04 bd c0 ac 35 c0 85 c0 74 05 89 43 28 89 c2 <8b> 32 85 f6 74 
28 89 34 24 e8 af 68 f3 ff 85 c0 0f 84 a7 00 00 
        <0>Kernel panic - not syncing: Attempted to kill init!

Himanshu Raj
PhD Student, GaTech (www.cc.gatech.edu/~rhim)
I prefer to receive attachments in an open, non-proprietary format.

Xen-users mailing list

<Prev in Thread] Current Thread [Next in Thread>