Re: An amusing bug. But what on earth should it be doing.

Walker, Bruce J (HP-Labs) wrote:
> Just looking at the code you sent, the unlock is dependent on id > 0  and id is a param to ssi_shm_cleanup so it is possible that ssi_shm_cleanup is called with id > 0 if a spinlock was already acquired and otherwise not.
>   
Yeah, but this is a cross-node IPC.

There should be no way one node could take a spinlock on another node in
one IPC and release it in another.



-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone