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-bugs] [Bug 1632] New: Dom0 crash if PCI passthrough VM crash or des

To: xen-bugs@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-bugs] [Bug 1632] New: Dom0 crash if PCI passthrough VM crash or destroy
From: bugzilla-daemon@xxxxxxxxxxxxxxxxxxx
Date: Wed, 30 Jun 2010 06:27:46 -0700
Delivery-date: Wed, 30 Jun 2010 06:27:58 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-bugs-request@lists.xensource.com?subject=help>
List-id: Xen Bugzilla <xen-bugs.lists.xensource.com>
List-post: <mailto:xen-bugs@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-bugs>, <mailto:xen-bugs-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-bugs>, <mailto:xen-bugs-request@lists.xensource.com?subject=unsubscribe>
Reply-to: bugs@xxxxxxxxxxxxxxxxxx
Sender: xen-bugs-bounces@xxxxxxxxxxxxxxxxxxx

           Summary: Dom0 crash if PCI passthrough VM crash or destroy
           Product: Xen
           Version: unstable
          Platform: x86-64
        OS/Version: Linux-2.6
            Status: NEW
          Severity: major
          Priority: P2
         Component: 2.6.18 dom0
        AssignedTo: xen-bugs@xxxxxxxxxxxxxxxxxxx
        ReportedBy: paraic.gallagher@xxxxxxxxx

Xen version 3.4.2
Kernel 2.6.18-164
Hardware IBM x3550 M2

If a VM using pci passthrough crashes or is destroyed by running 'xm destroy'
it can result in a Dom0 crash.

Output of dmesg on Dom0 after running 'xm destroy' shows something like:
ACPI: PCI interrupt for device 0000:10:00.1 disabled
irq 22: nobody cared (try booting with the "irqpoll" option)
Call Trace:
 <IRQ>  [<ffffffff802b36e6>] __report_bad_irq+0x30/0x7d
 [<ffffffff802b391d>] note_interrupt+0x1ea/0x22b
 [<ffffffff802b2e15>] __do_IRQ+0xbd/0x103
 [<ffffffff8029008e>] _local_bh_enable+0x61/0xc5
 [<ffffffff8026df5e>] do_IRQ+0xe7/0xf5
 [<ffffffff803af7d5>] evtchn_do_upcall+0x13b/0x1fb
 [<ffffffff802608d6>] do_hypervisor_callback+0x1e/0x2c
 <EOI>  [<ffffffff802063aa>] hypercall_page+0x3aa/0x1000
 [<ffffffff802063aa>] hypercall_page+0x3aa/0x1000
 [<ffffffff8026f501>] raw_safe_halt+0x84/0xa8
 [<ffffffff8026ca7c>] xen_idle+0x38/0x4a
 [<ffffffff8024b0c8>] cpu_idle+0x97/0xba
 [<ffffffff80646b05>] start_kernel+0x21f/0x224
 [<ffffffff806461e5>] _sinittext+0x1e5/0x1eb

[<ffffffff803e39be>] (usb_hcd_irq+0x0/0x55)
[<ffffffff803e39be>] (usb_hcd_irq+0x0/0x55)
[<ffffffff803e39be>] (usb_hcd_irq+0x0/0x55)
Disabling IRQ #22

I've tried booting with irqpoll options but this doesn't help.

The IRQ (normally IRQ 20 or IRQ 22 in testing) is used by another device such
as the RAID controller or USB hub, so the box stops functioning correctly. USB
device failure appears benign until you need to connect any USB device. If the
RAID controller is disabled usually the box stops responding immediately. 

A reboot is required to resolve the issue.
This occurs on the IBM x3550 M2 (S8800) with both the Intel and Broadcom cards
acting as the passthrough device.

If the guest is stopped using the 'xm shutdown' command no problems seem to

Configure bugmail: 
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

Xen-bugs mailing list

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