I have integrated your changes, and moved this bit referenced below, so it works with my
setup. Could you please test with yours, and make sure I didn't break anything?
configure.in | 13
examples/domain-events/events-c/event-test.c | 2
include/libvirt/virterror.h | 1
po/POTFILES.in | 1
python/libvir.c | 3
src/Makefile.am | 5
src/util.c | 4
src/virterror.c | 3
src/xen_inotify.c | 453 +++++++++++++++++++++++++++
src/xen_inotify.h | 31 +
src/xen_unified.c | 200 +++++++++++
src/xen_unified.h | 61 +++
src/xm_internal.c | 198 +++++++----
src/xm_internal.h | 6
src/xs_internal.c | 431 +++++++++++++++++++++++++
src/xs_internal.h | 51 +++
16 files changed, 1393 insertions(+), 70 deletions(-)
Daniel P. Berrange wrote on 11/21/2008 02:07 PM:
On Fri, Nov 21, 2008 at 01:56:16PM -0500, Ben Guthro wrote:
> I think I found the culprit of this:
>
> Daniel P. Berrange wrote on 11/21/2008 11:13 AM:
>> - Allocate capabilities info before initializing inotify driver
>> because loading XM config files /etc/xen requires this
>>
> ...
>> - if (!(priv->caps = xenHypervisorMakeCapabilities(conn))) {
>> - DEBUG0("Failed to make capabilities");
>> - goto fail;
>> - }
> If I undo this part of your patch, everything else works as designed.
>
> While I have no evidence yet to support this claim, I think this is
> failing trying to open the xend socket, perhaps before the xend driver
> is open...
Doh. Merge error. Originally I had moved it to be immediately before
the call to initialize xen inotify driver. For inexplicable reasons,
in the new patch I have it before the XenD driver.
Daniel