Ewan,
I have left two things open.
- I was not sure how to integrate with the vbd and vdi code that has
come in. These sections of code were commented out with
VBD<->VDI issue.
I don't know what this means. Could you explain?
The earlier code just used a field called image to refer to the disk
image. That field has been moved to the vdi_record_opt structure, I
think. I haven't yet written the code to read this structure.
- The upstream has not yet accepted the model info for vifs
patch that
Jim sent out.
Jim said that he would go and try a different approach
(putting the model info in a VIF.other_config dictionary), but
he hasn't yet come back to me with a new patch.
Cool.
This is just the first cut. I was not able to rigorously
test it since
my hvm box is still under the 'old' xen. I am hoping that by sending
it out, we will have more people working on it. With this, I can do
some basic enumeration etc.
The big change between 14542 and the tip seems to be some change in
the session handling code.
Calls on the session class now take two session references --
the first is the session that you're authenticating with, as
with any call, and the second is the session that you actually
want to talk about. This will allow us (later) to have the
ability for superusers to inspect other people's sessions, for example.
As far as you're concerned, just passing in the same session
handle twice should be all you need to do.
Thanks
Raj
4 messages in com.xensource.lists.xen-cimRE: [Xen-cim] [PATCH]move to changese...
.patch