On Thu, 2007-05-31 at 13:16, Sean Hefty wrote: > > Do you have some pointer/doc related to the design of current SA_CACHE > > module....It will make things faster to understand........if not then > > I will require your support to understand the things, Though I have > > some top level view. > > I don't have any design docs. But I will happily answer any questions. > > > Ok you mean Its not required to create a separate device interface in > > cache module as such. I think this is a good idea......Just for > > confirmation...whether /dev/umad is active on each node other than SM > > node? > > After giving this more thought, I like this approach. If we use a > vendor/application specific MAD class that used the SA class as a > template, we can begin creating a distributed SA. > > I haven't worked through details, but as an example, to load a PR into > the local SA, you could send it a 'Set' MAD with a PR in the data > portion. To load multiple paths, we could add a 'SetTable' method. To > remove a path, we would send a 'Delete' MAD. > > Whether or not the MADs are sent from the local node, or some other node > wouldn't matter.
Would there be some sort of weak authorization needed to do this (like some key of some sort) ? -- Hal > We can use this mechanism to pre-load the cache or > simply push updates to it. > > - Sean > _______________________________________________ > general mailing list > [email protected] > http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general > > To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
