|  |  | 
  
    |  |  | 
 
  |   |  | 
  
    |  |  | 
  
    |  |  | 
  
    |   xen-cim
RE: [Xen-cim] Patch for review 
| FYI - using upcalls to the CIMOM to
retrieve *all* instances of the association's target class (and subclasses)
and then filtering based on matching property values with source instance
is a nice generic way to implement associations that 'just works', and
for me facilitated rapid prototyping of working CIM providers for all the
numerous association classes. However, its not terribly efficient; and
once we start dealing with hundreds of DomU's it may become a run-time
problem for some of the associations.
 
 Just a heads-up that this is something
that we should consider optimizing where appropriate. I certainly focussed
on exposing functionality and minimizing bugs rather than optimizing for
efficiency when writing the original IBM providers.
 
 - Gareth
 
 Dr. Gareth S. Bestor
 IBM Linux Technology Center
 M/S DES2-01
 15300 SW Koll Parkway, Beaverton, OR 97006
 503-578-3186, T/L 775-3186, Fax 503-578-3186
 
 
 
 
 
| "Subrahmanian, Raj"
<raj.subrahmanian@xxxxxxxxxx> Sent by: xen-cim-bounces@xxxxxxxxxxxxxxxxxxx
 06/01/2006 09:48 AM
 | 
| To | "Jim Fehlig" <jfehlig@xxxxxxxxxx> |  
| cc | xen-cim@xxxxxxxxxxxxxxxxxxx |  
| Subject | RE: [Xen-cim] Patch for review |  
 
 |  
 
 Jim,
 I thought I had fixed this.
 I will take a look.
 Thanks
 Raj
 
 -----Original Message-----
 From: xen-cim-bounces@xxxxxxxxxxxxxxxxxxx
 [mailto:xen-cim-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Jim Fehlig
 Sent: Wednesday, May 31, 2006 9:34 PM
 To: Szymanski, Lukasz K
 Cc: xen-cim@xxxxxxxxxxxxxxxxxxx
 Subject: Re: [Xen-cim] Patch for review
 
 Szymanski, Lukasz K wrote:
 
 > Hello All -
 >
 > This patch adapts Raj's Xen_HostedComputerSystem patch to other
 > Xen_Hosted* files:
 >
 > Xen_HostedDisk.mof, Xen_HostedDisk.c, Xen_HostedProcessor.mof,
 > Xen_HostedProcessor.mof, Xen_HostedMemory.mof, Xen_HostedMemory.c,
 > Xen_HostedNetworkPort.mof, Xen_HostedNetworkPort.c
 >
 
 The problem with this patch, and Raj's previous one apparently :-), is
 that there is no filtering of object paths returned via upcalls to cimom
 for instance names of the associated thing.  For example, in
 ReferenceNames() in Xen_HostedMemory.c we make an upcall to get all
 instances of CIM_Memory when it is the target class.  But Xen_Memory
is
 a subclass of CIM_Memory so you get instances of Xen_Memory (which is
 the source class) and associate it with itself or other instances of
 Xen_Memory for other domains.  If I list associations of Xen_Memory
for
 domain 1 using this patch, I get a Xen_HostedMemory association for each
 Xen_Memory object produced by other domains.
 
 Make sense?  I'm quite tired now so this may be a smoked explanation
 :-).
 
 Thanks,
 Jim
 
 
 _______________________________________________
 Xen-cim mailing list
 Xen-cim@xxxxxxxxxxxxxxxxxxx
 http://lists.xensource.com/xen-cim
 
 _______________________________________________
 Xen-cim mailing list
 Xen-cim@xxxxxxxxxxxxxxxxxxx
 http://lists.xensource.com/xen-cim
 
 
 _______________________________________________
Xen-cim mailing list
Xen-cim@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-cim
 | 
 |  | 
  
    |  |  |