[ofa-general] hang on close in umem_release

[Pete Wyckoff]

rdreier at cisco.com wrote on Thu, 21 Jun 2007 08:40 -0700:
> hmm, I see what seems to be an "i can't believe it ever worked" type
> bug -- if the context is closing but then we do manage to get the mm's
> rwsem, it seems like we immediately try to lock it again, which
> obviously deadlocks.
> 
> Does this patch fix your problem and look correct?

Looks obviously correct and tests okay.  Ctrl-c in any situation
does the right thing now.  Before your refactoring of ib_umem, the
older version of ib_umem_release_on_close() did not have this
trylock optimization.  This new buggy code appears not to have shown
up in any releases yet, fortunately.

Thanks for the quick fix.

		-- Pete