EA builds with changes to object monitor implementation to avoid pinning with virtual threads
Yuval Lombard
yuval.l at securithings.com
Wed Apr 24 07:51:37 UTC 2024
Hey team!
I wanted to share some virtual threads tests output on build *23-loom+2-48.*
Pinning still occurs in both LockingMode=1/2.
Attaching some jdk.VirtualThreadPinned events caught by jfr.
*Lots of them:*
jdk.VirtualThreadPinned {
startTime = 10:31:19.654 (2024-04-24)
eventThread = "virtual-thread-65" (javaThreadId = 177, virtual)
stackTrace = [
java.lang.Object.wait0(long)
java.lang.Object.wait(long) line: 381
org.apache.hc.core5.concurrent.BasicFuture.get(long, TimeUnit) line: 101
org.apache.hc.core5.pool.StrictConnPool$1.get(long, TimeUnit) line: 184
org.apache.hc.core5.pool.StrictConnPool$1.get(long, TimeUnit) line: 178
...
]
}
jdk.VirtualThreadPinned {
startTime = 10:40:24.750 (2024-04-24)
eventThread = "virtual-thread-209" (javaThreadId = 338, virtual)
stackTrace = [
java.lang.Object.wait0(long)
java.lang.Object.wait(long) line: 381
org.eclipse.paho.client.mqttv3.internal.Token.waitForResponse(long)
line: 145
org.eclipse.paho.client.mqttv3.internal.Token.waitForCompletion(long)
line: 108
org.eclipse.paho.client.mqttv3.MqttToken.waitForCompletion(long) line:
67
...
]
}
I am well aware that java.lang.Object.wait continues to pin, but wanted to
point out a few of the 3rd party libs that are used so you get the
impression of the necessity and the anticipation for the more robust
solution :)
--
Kind regards,
*Yuval Lombard*
*Lead Software Engineer*
+972.50.548.0111
yuval.l at securithings.com
[image: logo_black.png]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/loom-dev/attachments/20240424/3f20af6d/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: logo_black.png
Type: image/png
Size: 99833 bytes
Desc: not available
URL: <https://mail.openjdk.org/pipermail/loom-dev/attachments/20240424/3f20af6d/logo_black-0001.png>
More information about the loom-dev
mailing list