[openib-general] [PATCH] IB/ucm: fix deadlock in cleanup

[Michael S. Tsirkin]

> >I seem to have lost this, maybe because I was waiting for an ack from
> >you, and also a report back from Michael about whether it helped.  Do
> >you guys both agree that this is a correct fix?
> 
> This fix definitely looks like it's needed to prevent a potential deadlock, but
> I haven't seen anything from Michael about the results of any tests.

There actually was an Ack from Sean, but I forgot that I promised to post the results
of my tests, so I kind of hoped it would make it in 2.6.19.

It does actually seem to fix a deadlock we see in regression - so I
think if not 2.6.19 we'll want it for -stable.

-- 
MST