-
Bug
-
Resolution: Fixed
-
P3
-
9
-
b24
| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-8216218 | 13 | Erik Gahlin | P3 | Resolved | Fixed | team |
| JDK-8217834 | 11.0.4-oracle | Erik Gahlin | P3 | Resolved | Fixed | b01 |
| JDK-8217928 | 11.0.3-oracle | Erik Gahlin | P3 | Closed | Fixed | b05 |
| JDK-8219532 | 11.0.3 | Erik Gahlin | P3 | Resolved | Fixed | master |
| JDK-8240390 | openjdk8u262 | Erik Gahlin | P3 | Resolved | Fixed | team |
<value type="MILLIS" field="timeout" label="Park Timeout"/>
and the trace code copies the "time" jlong passed to Unsafe::park right into the timeout field in the event.
The problem is that the specification for Unsafe::park is a bit more complex than that.
* If the "isAbsolute" jboolean is true then the interpretation of the "time" parameter should be a "milliseconds since epoch" timestamp.
* If "isAbsolute" is false then the interpretation of the "time" parameter should be a timeout in nanoseconds relative to "now".
The first case appears to require the EPOCHMILLIS type whereas the second case requires the NANOS type so the current implementation is incorrect for any value of "isAbsolute"
- backported by
-
JDK-8216218 Trace event for thread park has incorrect unit for timeout
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Closed
-
- relates to
-
JDK-8239140 Backport JFR to OpenJDK 8
-
- Resolved
-