[ewg] Re: pkey change handling patch

[Roland Dreier]

 > 1. Direct access in ib_find_pkey will probably heart RC connections
 > per second rate.

I think it's probably in the noise.  And anyway I don't think the
connection rate of ipoib CM is particularly important.  And we can
always optimize ib_find_pkey as part of the longer-term plan to get
rid of ib_find_cached_pkey.

If you want to tackle more of the cache elimination plan we discussed
that would be great though.

 > 2. What do you think about OrG's opinion (I'm copying it from the other thread):

He seems to be saying that it's OK to introduce a window where things
fail spuriously.  I disagree.