Well, the critical thing is that on a large fabric, capturing all the MADs 
would be quite a firehose of data. Being able to filter on a single source or 
destination would make the amount of data more manageable.

I'm flexible though, if the act of dumping all the data to a user-space app 
doesn't adversely affect the system too much, we could always have the 
user-space tool do the filtering.

-----Original Message-----
From: Hefty, Sean [mailto:[email protected]] 
Sent: Thursday, October 14, 2010 12:03 PM
To: Mike Heinz; Linux RDMA list
Subject: RE: [RFC 1/2] IB/mad: Simplify snooping interface

> Would it be worthwhile to filter based on the source LID/GID of the MAD?

Without breaking the umad ABI, GID filtering would be difficult.  LID is 
possible, but I can't think of a good reason why the kernel should perform that 
level of filtering.  Ideas?

I have thought about using the umad_send interface to specify more complex 
filters that could be handled by ib_user_mad, rather than ib_mad.  I just 
didn't want to get too fancy in the kernel, especially for an initial 
submission.

- Sean 

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to