RFR(s): 8023541 Race condition in rmid initialization
david.holmes at oracle.com
Wed Jan 29 09:05:13 UTC 2014
This looks fine to me.
Tristan: the initialized field is only accessed under synchronization so
does not need to be volatile.
On 29/01/2014 4:51 PM, Stuart Marks wrote:
> Hi all,
> Please review this fix to a race condition in rmid initialization.
> Briefly, rmid subclasses the RMI registry implementation and provides
> special handling for its own stub. Unfortunately the registry is
> exported in the super() call, making remote calls possible before rmid's
> stub initialization is complete. The fix is to ensure that all remote
> calls wait for initialization before proceeding.
More information about the core-libs-dev