Loading...

XML

Word

Printable

Type: Bug
Resolution: Fixed
Priority: P3
Fix Version/s: 15
Affects Version/s: 15, 16
Component/s: hotspot
Labels:
- oracle-triage-15

Subcomponent:
jfr
Resolved In Build:
b31

Issue	Fix Version	Assignee	Priority	Status	Resolution	Resolved In Build
JDK-8248992	16	Markus Grönlund	P3	Resolved	Fixed	b05
JDK-8250148	15.0.2	Markus Grönlund	P3	Resolved	Fixed	b01
JDK-8250447	15.0.1	Markus Grönlund	P3	Resolved	Fixed	b03

This bug will attempt to resolve the problem described in ~~JDK-8243962~~.

The bug fix complements the work done in ~~JDK-8245042~~ and ~~JDK-8245113~~ to avoid having a JavaThread that is reading OS performance counters, which has proven to be very slow on certain platforms and circumstances, stall a pending safepoint synchronization.

Transitions will be performed for periodic events EventCPULoad and EventThreadContextSwitchRate.

backported by

JDK-8248992 Transition JFR Periodic Task Thread to "_thread_in_native" before invoking performance counters

Resolved

JDK-8250148 Transition JFR Periodic Task Thread to "_thread_in_native" before invoking performance counters

Resolved

JDK-8250447 Transition JFR Periodic Task Thread to "_thread_in_native" before invoking performance counters

Resolved

relates to

JDK-8243962 Various JVM TI tests time out using JFR on Windows

Resolved

JDK-8245042 Improve scalability of reading Windows Performance counters via PDH when using the Process object

Resolved

JDK-8245113 JFR Recorder Thread to run in thread state "_thread_in_native"

Resolved

(1 relates to)

Assignee:: Markus Grönlund

Reporter:: Markus Grönlund

Votes:: 0 Vote for this issue

Watchers:: 2 Start watching this issue

Created:: 2020-07-03 03:06

Updated:: 2024-10-16 19:48

Resolved:: 2020-07-07 08:59

Details

Backports

Description

Attachments

Issue Links

Activity

People

Dates