Re: [Xen-devel] [BUG] Xen BUG at irq.c:1705 after [VT-D]d1:PCIe: unmap

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 05.03.16 at 13:22, <hakon@xxxxxxxxxxxxxxxxxx> wrote:
> This has happened in 2 out of four recent shutdowns of a specific domU
> "garbo.hvm", domid=1.
> ------- domU specs (no tmem in cmdline arg to domU): -----
> Linux garbo 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u4
> (2016-02-29) x86_64
> garbo:~# cat /proc/cmdline
> BOOT_IMAGE=/boot/vmlinuz-3.16.0-4-amd64
> root=UUID=042054cf-2a70-46de-bd90-c712fbedbf81 ro intel_iommu=on 3
> console=ttyS0 console=vga
> --------
> serial log:
> (XEN) [2016-03-04 23:38:48] tmem: allocating persistent-private tmem
> pool for domid=3...<G><2>pool_id=0
> (XEN) [2016-03-04 23:39:12] grant_table.c:1491:d2v1 Expanding dom (2)
> grant table from (5) to (6) frames.
> (XEN) [2016-03-04 23:40:28] [VT-D]d1:PCIe: unmap 0000:08:00.0
> (XEN) [2016-03-04 23:40:28] [VT-D]d0:PCIe: map 0000:08:00.0
> (XEN) [2016-03-04 23:40:30] tmem: flushing tmem pools for domid=1
> (XEN) [2016-03-04 23:40:30] Destroying persistent-private tmem pool
> domid=1 pool_id=0
> (XEN) [2016-03-04 23:40:30] Xen BUG at irq.c:1705
> (XEN) [2016-03-04 23:40:30] ----[ Xen-4.6.1  x86_64  debug=y 
> Tainted:    C ]----

Plain Xen 4.6.1 doesn't have any BUG() or BUG_ON() at that line,
so you will need to tell us which nearby one it is. And of course
the question then also arises whether plain 4.6.1 (or even better
4.7-unstable) would also surface this problem.

> I do have qemu-dm logs for the domain, but I'm not certain which log
> goes with which run. Attaching what I THINK are correct. My timezone is
> GMT+1, xen seems to be at GMT, so qemu-dm with file-system timestamp Mar
> 5 00:40 ought to be right.

Not how the domain ID is being logged at the very top of the file,
to aid making such association. (But yes, from the looks of it you
picked the right one.)

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel