| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-8083017 | emb-9 | Axel Siebenborn | P4 | Resolved | Fixed | team |
| JDK-8086479 | 8u65 | Axel Siebenborn | P4 | Resolved | Fixed | b01 |
| JDK-8076297 | 8u60 | Axel Siebenborn | P4 | Resolved | Fixed | b12 |
| JDK-8137573 | emb-8u65 | Unassigned | P4 | Resolved | Fixed | b01 |
| JDK-8078928 | emb-8u60 | Axel Siebenborn | P4 | Resolved | Fixed | team |
I investigated a crash in ClassLoaderDataGraphKlassIteratorAtomic::next_klass_in_cldg where class is 0. This can happen as _next_klass is changed concurrently, but is not correctly qualified as volatile.
The declaration volatile Klass* _next_klass indicates that the Klass object may change, but actually the pointer is changed.
I have seen the crash on linux s390x
The declaration volatile Klass* _next_klass indicates that the Klass object may change, but actually the pointer is changed.
I have seen the crash on linux s390x
- backported by
-
JDK-8076297 Wrong volatile qualifier for field ClassLoaderDataGraphKlassIteratorAtomic::_next_klass
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-