New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Swing Application crashes / hangs since update from 8.275.01.1 to 8.282.08.1 #305
Comments
|
Hi @nrmnrm Thank you for reaching out, We are investigating this issue and we will keep you updated. |
Thanks, let me know if I can do any testing with a new version as we cen reproduce it rarther quickly. |
Hi @nrmnrm , It's hard to know for certain if this build will fix your issue, However, Please let us know if this works. amazon-corretto-8.292.10.1-windows-x86-jdk.zip test build In addition, does the JVM always crash at |
Unfortunately it still either freezes or closes. Not sure where the freezing occurs, the closing happend again in the layoutMajorAxis area. |
|
Thank you for testing the build. Would it be possible to share a reproducer for this issue?. It would speed up the investigation on our end. |
You mean a video? Otherwise its a really complex client, server, DB setup.. |
Could you isolate the UI failure into a separate program and share that?. |
That would mean a lot of effort, it is a really huge swing application. I was hoping the difference to the working AdoptOpenJDK would help. Must be something from the 8.282.08.1 build. |
Could you try running with this jvm parameter?
|
I guess you mean (double X) Yes, that seems to fix it as I am not able to reproduce |
Could you try this |
Now I get this from time to time: there was a mdmp file created (500mb), need that too? I removed that -XX flag again for that test, correct?! |
When you removed -XX:CompileCommand='exclude,javax.swing.text.BoxView::layoutMajorAxis', you application again crashed at C2-compiled layoutMajorAxis. That's a sign that c2 has problem on that method on Windows x86. If the workaround works for you, we believe it has unblocked you, right? If you'd love to contribute openjdk and Corretto, you can help us in any of those ways.
|
Hi, regarding "If the workaround works for you, we believe it has unblocked you, right?": |
Hello, @nrmnrm This query shows the difference between 8.275.01 and 8.282.08.1 / No big change made in c2 compiler, however, we did upgrade vc++ in this commit. |
hi, @nrmnrm, I check our CI records. c34e5b8 passed all tests like others. Now my guess is that it may run faster than prior binary and happens to unveil a bug in C2. I can confirm that there's no difference of C2 between 8.275.01.1 and 8.282.08.1 It's hard to say why other distributions don't have this issue. We still need your help to pinpoint the root cause. Are you a subscriber of AWS premium support? If so , please contact your customer support representatives. They will provide you a channel to share binaries with us. If we get a reproducible, it's easier to analyze the problem. |
Hi, would you be able to build with your old vc++ environment that you were using before 8.282.08.1 and let me test that? Thanks and best regards |
Hi, just checking again, would it be possible to build with the older VC++ environment so i could test that? |
Is this still an issue with the most recent release? Have you tried JDK11? |
we cant switch to JDK11 unfortunately if there were any changes to latest build that could have fixed the issue i can retest!? EDIT: i just retested, still freezes immediately (currently I cannot reproduce complete crashes, so I can not provide debug info as requested above via XX:CompileCommand) Additionally I found out that the workaround So 8.275.01.1 is the last 32bit Java 8 that works for us. |
I have created a dump file of the hanging application, it is compressed around 100mb which I cannot upload here, where can I put it for you? |
Could you post a new link to a fastdebug build? I'm trying to sort through another Swing crash that has similar symptoms, but a different stack trace. |
hi, @nrmnrm, Application freezing may be a different issue from c2-generated code crash like you post above. When it happens, you can use external tool such as jstack to investigate why it is hanging.
If your application contains sensitive data in memory, please don't share the coredump file on the internet. |
A mentioned in my initial report: and there is no hanging or closing with 8.275.01.1 stack is from 8.302 now |
@nrmnrm, |
@navyxliu yes, I can confirm -XX:-UseFPUForSpilling fixes it! |
Hi,
we found a bug that appears to be corretto specific since 8.282.08.1 (32bit for Windows).
The build https://github.com/AdoptOpenJDK/openjdk8-binaries/releases/download/jdk8u282-b08/OpenJDK8U-jre_x86-32_windows_hotspot_8u282b08.zip is working fine corretto 8.282.08.1 is not
Unfortunantely it is hard for you to reproduce. When we are within our rich client swing application the client hangs or closes completely from time to time. This seems to be a timing issue, so you have to click around the application rather quickly. We could provide a video if needed. But maybe you already have a suspect candidate when you compare
Corretto 8.282.08.1 with AdoptOpenJDK jdk8u282-b08 where this is not happening.
Latest Corretto build 8.292.10.1 has the same issue
hs_err_pid65924.log is attached (complete crash of jvm)
The text was updated successfully, but these errors were encountered: