-
Sub-task
-
Resolution: Fixed
-
P3
-
24
-
b06
| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-8347575 | 24 | Christian Hagedorn | P3 | Resolved | Fixed | b32 |
The Java Fuzzer found a wrong execution which could be traced back to JDK-8325495 which went into JDK 24. Since we are close to RDP 2, a backout seems the safest option for JDK 24. A REDO can be done for JDK 25.
Reproducers fails intermittently (~20%):
Reduced.java:
$ for i in $(seq 1 20); do echo $i; java -XX:CompileCommand=quiet -XX:CompileOnly=*Reduced*::* -Xbatch -XX:+StressIGVN -cp ~/javafuzzer Reduced.java; done
Failure:
Exception in thread "main" java.lang.RuntimeException: Expected: 982, was: 198
at Reduced.main(Reduced.java:20)
Test.java is the original Java Fuzzer test. Run it with the same flags as for Reduced.java and compare C2 compiled version against run with -Xint.
Reproducers fails intermittently (~20%):
Reduced.java:
$ for i in $(seq 1 20); do echo $i; java -XX:CompileCommand=quiet -XX:CompileOnly=*Reduced*::* -Xbatch -XX:+StressIGVN -cp ~/javafuzzer Reduced.java; done
Failure:
Exception in thread "main" java.lang.RuntimeException: Expected: 982, was: 198
at Reduced.main(Reduced.java:20)
Test.java is the original Java Fuzzer test. Run it with the same flags as for Reduced.java and compare C2 compiled version against run with -Xint.
- backported by
-
-
- Resolved
-
- relates to
-
JDK-8325495 C2: implement optimization for series of Add of unique value
-
- Closed
-
- links to
-
Commit(jdk24)
openjdk/jdk/f42e2c10
-
Commit(master)
openjdk/jdk/062f2dcf
-
Review(jdk24)
openjdk/jdk/23077
-
Review(master)
openjdk/jdk/23066
(1 links to)