| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-8083219 | emb-9 | Filipp Zhinkin | P4 | Resolved | Fixed | team |
One of compiler/rtm/locking/TestRTMLockingThreshold testcases verifies that method will be recompiled on low abort ratio w/ LockingThreshold=0.
Test artificially abort every second transaction, so with LockingThreshold=0, first abort will happened on second transaction. Test verifies that there is exactly two transaction's and one abort.
However, abortion may happen during first transaction due to h/w interrupt. In this case method will be recompiled after first abort as it should, but test will fail because there will be only one transaction.
This issue is very intermittent and reproduces ~ once per 1000 runs.
Test artificially abort every second transaction, so with LockingThreshold=0, first abort will happened on second transaction. Test verifies that there is exactly two transaction's and one abort.
However, abortion may happen during first transaction due to h/w interrupt. In this case method will be recompiled after first abort as it should, but test will fail because there will be only one transaction.
This issue is very intermittent and reproduces ~ once per 1000 runs.
- backported by
-
JDK-8083219 compiler/rtm/locking/TestRTMLockingThreshold test may fail if transaction was aborted by interrupt
-
- Resolved
-
- is blocked by
-
JDK-8050486 compiler/rtm/ tests fail due to monitor deflation at safepoint synchronization
-
- Resolved
-