[openib-general] question regarding GRH flag in ib_ah_attr

[Hal Rosenstock]

On Thu, 2006-05-11 at 13:12, Jason Gunthorpe wrote:
> On Thu, May 11, 2006 at 07:20:19AM -0400, Hal Rosenstock wrote:
> 
> > That would be a simpler check but HopLimit is not a required component
> > of PathRecord but I think this may not be sufficient as just because a
> > HopLimit >= 2 doesn't mean that a packet would be forwarded off subnet.
> 
> I was thinking of the other direction: How does the requestor/client
> know if a Path requires a GRH.

The requester/client needs to request a path for a DGID which is off
(the local) subnet.

> To allow what Roland is talking about you need an unambiguous
> mechanism where the SA can signal to the client that the path
> needs a GRH.

Ah, you are referring to the SA path record response not the request.

> The only field I can see that could be used for that is HopLimit..

That's one. The ugly prefix comparison would be another.

> Think of it the other way, HopLimit < 2 means it _can't_ be forwarded
> off subnet, so that result from the SA should _always_ cause the
> requesting client to not use a GRH for that path.

Not always true in terms of local subnet (multicast and management MAD
response exceptions).

> Any test beyond HopLimit could be done in the SA prior to returning
> the path records to the client.

Are you saying HopLimit is supplied to the SA in the request ? It could
be but it's optional in general. In the router case, an off subnet DGID
should be sufficient. I would think the HopLimit (as well as the other
GRH fields) would need to be returned by the SA to the client.

> If further tests are put in the client
> they only limit the routing configurations that are possible.

Not sure what further tests you are referring to here. I agree with the
goal not to add any unnecessary constraints on routing configurations.

> Note:
> Although 8.3.6 specifies that 0 and 1 don't let the packet off
> the subnet table 60 says that CA's should set the HopLimit
> to 0 and the 'first' router should fill it in. Hmm..

Interesting. The description is table 60 also says "Alternately set
according to application."

> > Why is a request with just a non link local prefix (with HopLimit
> > wildcarded) not sufficient ?
> 
> I think it wouuld be best of the SA had full control over what headers
> the CA's put on their packets on a path by path basis. That allows for
> the most flexability down the road.

Not sure exactly what you mean by full control over the routing header
(GRH). The SA supplies the info for the headers to the client and the
client is responsible for putting the correct info in the headers. Do
you mean supplies sufficient info for the client to do this correctly ?
If so, I agree.

-- Hal

> Jason