[ofa-general] Re: IPoIB path caching

[Sean Hefty]

> When I said "requires" I meant that "I think that it is required", I 
> agree that the current IPoIB code does not link them together. In the 
> beginning on this thread Roland commented saying he agree with me, but 
> since then he did not provide more input to the discussion...

I think Jason said it clearer.  Think of the local SA as part of a 
limited, distributed SA, rather than a cache.

> what's the reasoning to update the cache when there is GID out of 
> service? you don't cache IP 2 GID mapping, nor the IB SA provides GID 
> resolution services... What do you do on GID OUT, just remove all the 
> paths in the cache for which this is the DGID?

Paths associated with that GID are removed.  When the GID comes back 
into service, a new query is issued for just that GID, and the resulting 
paths are re-added.  Whether the local SA responds to GID in/out of 
service is run-time configurable.  (An administrator may not want the 
local SA to sync during some application runs, to avoid the traffic or 
hit on the SA.)

> Does it mean that you will have to re-implement RMPP in a user space 
> library or just the initiation of the query would be from user space?

RMPP would not need to be re-implemented, but I haven't come up with a 
user space solution that I like either.  I'm not trying to design a full 
blown distributed SA, nor do I want to try to standardize some new SA 
protocol.  I only want to fix the SA PR scalability problems that we see 
in practice.

- Sean