[openib-general] Re: [PATCH] cm refcount race fix

Roland Dreier rdreier at cisco.com
Mon May 8 09:47:32 PDT 2006


    Michael> I do lock/unlock *after* refcount is 0: after wait_event.
    Michael> Please tell me what I am missing.

Sorry, I misread your patch and then miscopied it somehow.
You're right, it would work.

But I think changing atomic_t to an integer protected by a lock it
much cleaner anyway.

 - R.



More information about the general mailing list