Pkg xen devel - Feb 2012 - Boot process hangs then reboots when using Xen + Linux 3.2

Hi,

I just upgraded to the latest SID after reinstalling my test server.
Then, after a while, during the boot process, Xen seems to hang, then my
server reboots.

It's difficult for me to see anything, since I'm using a KVM over IP,
using a quite slow link (from China to UK, using a FreeNX box that
connects through a Java KVM over IP to the server), and I don't think
it'd be easy for me to setup a serial console.

I tried to run kernel 3.1 from snapshot.debian.org (in fact, 3.1.8-1),
and I had no issue. Running the latest 3.2 from SID without Xen isn't an
issue either (eg: it boots as expected).

Before I start use reportbug, and/or try to investigate more, I just
wanted to check: has anyone seen this issue in SID? Is this a known problem?

Cheers,

Thomas Goirand (zigo)

On 02/13/2012 09:28 PM, Pasi K?rkk?inen wrote:
> I haven't tried the latest packages, so I can't comment on that..
> 
> but it sounds like you need to set up a serial console (or Serial Over LAN)
to debug that:
> http://wiki.xen.org/wiki/XenSerialConsole
> 
> 
> -- Pasi
Thanks to Pasi's help on IRC (thanks a lot), I've been able to use the
SOL of my server, and sent it as a bug report for the kernel (see #660554).

I've attached the log file to this mail anyway, because it doesn't crash
without Xen (so it must be related somehow).

Cheers,

Thomas Goirand (zigo)
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: log.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120220/de47989a/attachment.txt>

On Mon, Feb 20, 2012 at 04:40:00AM +0800, Thomas Goirand
wrote:
> (XEN) [VT-D]dmar.c:528:   RMRR address range not in reserved memory base =
bf7ec000 end = bf7ebfff; iommu_inclusive_mapping=1 parameter may be needed.
> (XEN) [VT-D]dmar.c:585:   The RMRR (bf7ec000, bf7ebfff) is incorrect!
> (XEN) Failed to parse ACPI DMAR.  Disabling VT-d.
> (XEN) ERST table is invalid
Broken ACPI table, aka firmware bug.
> (XEN) mm.c:907:d0 Error getting mfn 3209e (pfn 5555555555555555) from L1
entry 000000003209e023 for l1e_owner=0, pg_owner=0
> (XEN) mm.c:4962:d0 ptwr_emulate: fixing up invalid PAE PTE 000000003209e023
> (XEN) mm.c:907:d0 Error getting mfn 3209f (pfn 5555555555555555) from L1
entry 000000003209f023 for l1e_owner=0, pg_owner=0
> (XEN) mm.c:4962:d0 ptwr_emulate: fixing up invalid PAE PTE 000000003209f023
This looks broken.
> [    3.691809] [Firmware Bug]: ERST: ERST table is invalid
Now the kernel detects the broken table.
> (XEN) ../physdev.c:155: dom0: wrong map_pirq type 3
Hmm.
> [    8.771023] invalid opcode: 0000 [#1] SMP 
> [    8.771031] Modules linked in: ioatdma(+) i2c_i801 serio_raw pcspkr dca
i7core_edac i2c_core joydev evdev processor button edac_core iTCO_wdt
iTCO_vendor_support thermal_sys ext3 jbd mbcache dm_mod raid10 raid1 md_mod
sd_mod crc_t10dif usbhid hid uhci_hcd ahci libahci libata scsi_mod ehci_hcd
usbcore usb_common e1000e [last unloaded: scsi_wait_scan]
> [    8.771131] 
> [    8.771136] Pid: 0, comm: swapper/0 Not tainted 3.2.0-1-686-pae #1
Supermicro X8STi/X8STi
> [    8.771148] EIP: 0061:[<f74df33d>] EFLAGS: 00010246 CPU: 0
> [    8.771156] EIP is at __cleanup+0xd6/0x113 [ioatdma]
To check what is really going on here, I need the _exact_ version of
the kernel.
> [    8.771505] Kernel panic - not syncing: Fatal exception in interrupt
Okay, this is not surprising.

Bastian

-- 
There are always alternatives.
		-- Spock, "The Galileo Seven", stardate 2822.3

On 03/02/2012 04:55 PM, Bastian Blank wrote:
> On Fri, Mar 02, 2012 at 01:20:17PM +0800, Thomas Goirand wrote:
>> Replacing BUG_ON by a WARN_ON, and adding #define DEBUG 1 on top of
>> dma_v2.c, my kernel booted, and I had the attached dmesg output.
> 
> Please provide /proc/interrupts from all four cases (32/64, xen/non-xen)
> and the Xen APIC-state (z on the Xen console I think).
> 
> Bastian
As per your request, see attached files.

Thomas
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: proc_interrupts_32bits_kernel.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120302/c7e43491/attachment-0006.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: proc_interrupts_32bits_kernel_no_xen.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120302/c7e43491/attachment-0007.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: proc_interrupts_64bits_kernel.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120302/c7e43491/attachment-0008.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: proc_interrupts_64bits_kernel_no_xen.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120302/c7e43491/attachment-0009.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: xen_acpi_state_32bits_kernel.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120302/c7e43491/attachment-0010.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: xen_acpi_state_64bits_kernel.txt
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20120302/c7e43491/attachment-0011.txt>