[ofa-general] Re: having problems with the latest for-2.6.27 tree

[Or Gerlitz]

Roland Dreier wrote:
> Would have been nice for the original author of a patch changing locking to test with lockdep (IB needed the mutex_lock_nested treatment too), but oh well.
Yes, sounds like testing the original patch with lockdep would have 
bumped something earlier, thanks for fixing that.
> yep, the mutex_lock for conn_id in iw_conn_req_handler should really be
>
> 	mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING);
>
> since it's already inside another lock of the same class.  but I don't
> understand why it leads to a real deadlock... they're different mutexes
> and I don't see how the conn_id lock could already be locked??
I am not sure to follow what does "class" means in this context, is this 
struct rdma_id_private? such that "inside another lock of the same 
class" means locking the same  mutex for  another instance of this 
structure?

Or.