(SPAM?) Re: [openib-general][PATCH][RFC]: CMA header

[Sean Hefty]

Guy German wrote:
>>> typedef void (*ib_cma_event_handler)(enum ib_cma_event event, void 
>>> *context, const void *private_data);
>>> typedef void (*ib_cma_listen_handler)(void *cma_id, struct ib_device 
>>> *device, void *private_data, void *context);
>>
>> I think we can merge these two handlers.  We do not want to pass back 
>> struct ib_device* to a caller.  The device needs to be associated with 
>> the cma_id up front.
> 
> The listen handler passes also the new cma_id. Do you think it is better 
> to merge them and pass null in this field in the "active side" case? (or 
> maybe pass the same cma_id back in the handler)...

It seems like something like:

ib_cma_event_handler(struct ib_cma_id*, struct ib_cma_event event);

struct ib_cma_event {
	enum ib_cma_event event;
	/* per event needed data, if any... */
	struct ib_cma_id *listen_id;
	void *private_data;
	int private_data_len; /* needed? */
}

would work for any case.  I don't know that we need to distinguish between 
active and passive sides once a connection is established, and if we want to 
support peer to peer connections at some point in time.  It's just not clear to 
me that once a connection has been established, if the listen handler is still 
invoked for the new cma_id (e.g. disconnect), or is the cma_event_handler invoked?

- Sean