| |
|
 |
|
 |
|
|
|
| |
|
|
xen-ia64-devel
[Xen-ia64-devel] Re: PATCH: xencomm
On Wed, 2006-09-27 at 10:59 +0200, Tristan Gingold wrote:
> Hi,
>
> here are 3 patches for xencomm.
>
> The first one (a2) is the xencomm in itself. Compatibility is broken.
> The second one (b2) renumbered dom0vp. Because compatibility is broken
> dom0vp
> can be renumbered safely.
> [Are there good reasons why the numbering was sparse ?]
> The last one (c2) is a small optimization for physdevop eoi.
>
> tested by booting domU with netbk,netloop and blkbk as modules in dom0.
Hi Tristan,
I'm seeing a couple problems. The first is this:
# xm top
(XEN) xencomm_copy_to_guest error: f0000040fda98000 magic was 0x0
Failed to retrieve statistics from libxenstat
The second is that if I put a domU (UP, 1GB RAM, w/ vnif) in a reboot
loop (the last init script for the final runlevel issues a reboot), I
can only get about 20-25 loops before I start getting these:
(XEN) lookup_domain_mpa: d 0xf000000007b30080 id 23 current 0xf000000007aa8000
id 0
(XEN) lookup_domain_mpa: non-allocated mpa 0x0 (< 0x40000000)
(XEN) lookup_domain_mpa: d 0xf000000007b30080 id 23 current 0xf000000007aa8000
id 0
(XEN) lookup_domain_mpa: non-allocated mpa 0x3a770000 (< 0x40000000)
(XEN) lookup_domain_mpa: d 0xf000000007b30080 id 23 current 0xf000000007aa8000
id 0
(XEN) lookup_domain_mpa: non-allocated mpa 0x3a770000 (< 0x40000000)
The first time I hit this, it led to an Alt DTLB fault in
vcpu_get_domain_bundle(). The second time, it just seems to be
infinitely printing lines like above. This is the first time I've tried
such an experiment, so it may not be related to the xencomm patch.
Removing the vnif configuration from the domU appears to allow the
reboot test to continue much longer. Any thoughts on these? Thanks,
Alex
--
Alex Williamson HP Open Source & Linux Org.
_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel
|
|
|
| |
|
Home