That seems reasonable Amanda. I've run the updated test for a while and
I can't see any problem with it.
Thanks,
Michael
On 10/02/14 08:56, Amanda Jiang wrote:
Hi All,
Please review the fix for JDK-8031661
http://cr.openjdk.java.net/~tyan/amandaj/JDK-8031661/webrev.01/
Description:
Root Cause:
In line 195 and 196, the two actions can be ordered by a
happens-before relationship, which will cause synchronization fail.
t1cond1latch.countDown();
t1cond2latch.await();
Suggested Fixes:
Using only one CyclicBarrier ("t1Cond1") sync c1,c2 and main thread
