[openib-general] another ipoib question

[Michael S. Tsirkin]

Hello, Roland!

Here's another question from code review conducted by Eitan Rabin:

could flush task set the ipoib_neigh pointer encoded inside the
neighbour hardware address to NULL and free the neighbour,
while ipoib_start_xmit is accessing the ipoib_neigh through
the pointer is has loaded from the hardware address?

flush task does not seem to hold xmit_lock.

-- 
MST