On Mar 26, 2008, at 1:42 PM, Andrew Doran wrote:
For sockets: I want to be able to do container based locking / "lockchaining" similar to our scheduler, where the locks are objects in theirown right. For processes, clone() is quite granular. It seems a lot easier to do it this way rather than inventing another data structure to hang off struct proc.
Is we are going to have one mutex / cacheline, should we throw a kcondvar_t
in there as well (even if we never use it)?