-
Notifications
You must be signed in to change notification settings - Fork 728
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
jdk19 OpenJDK java/lang/Thread/virtual/stress/Skynet.java hang / timeout #16728
Comments
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_sanity.openjdk_ppc64_aix_Nightly/110/
|
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_sanity.openjdk_ppc64le_linux_Nightly/111
|
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_sanity.openjdk_x86-64_linux_Nightly/111
|
Changes 9160108...711ca6c |
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_sanity.openjdk_x86-64_mac_Nightly/110
|
Likely caused by adoptium/aqa-tests#4337. @LinHu2016 can you confirm if the segfault is related to #16713 |
#16728 (comment)
Also seen at x86-64_mac |
The time out occurred at an internal build(
|
|
it is possible same issue with #16652, the issue(#16652) is not directly related with large cpus, but we have not seen the problem with OS native threads smaller than 64. |
Reopen until the backport to 0.37 is completed. |
The test needs to be unexcluded adoptium/aqa-tests#4354 |
Skynet is unexcluded via adoptium/aqa-tests#4491 |
In my local run, the JIT was fully enabled. Locally, I couldn't reproduce the error with |
as far as I can see, J9VMThread AIX specific fields:
all these fields are only relevant, when there are JIT-ed code running. (there is codertTOC field which seemed used by Z system only). And, I knew the latter 3 fields are not used at all nowadays (RI was disabled 3 years ago). |
No, I cannot think of one. |
What about if something is corrupted and the VM tries to return to JIT-ed code that doesn't actually exist. |
I think it would be good to enable some stack walking tracing. For example, I can see something like:
But I am not sure how to enable it. Maybe some other VM tracing during stack unwinding would be useful? |
Stack walk tracing is enabled via |
I've been running grinder [1] with [1] https://hyc-runtimes-jenkins.swg-devops.com/job/Grinder/37043/ |
Correction to the above: the tests did fail but due to a timeout. Running without |
Currently, recycled continuations are reset and initialized in freeContinuation at the end of their use. There are also gaps where the recycled continuation might retain old values and not be properly reset. The above approach leads to corruption and segfaults as seen in the runJavaThread failure reported in eclipse-openj9#16728. To fix the above crash, all continuations (new and recycled) are initialized and reset in createContinuation just before the continuation begins execution. Related: eclipse-openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
I was able to get a better native stack for the
|
Excellent! And just to confirm: with |
Currently, recycled continuations are reset and initialized in freeContinuation at the end of their use. There are also gaps where the recycled continuation might retain old values and not be properly reset. The above approach leads to corruption and segfaults as seen in the runJavaThread failure reported in eclipse-openj9#16728. To fix the above crash, all continuations (new and recycled) are initialized and reset in createContinuation just before the continuation begins execution. Related: eclipse-openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
Currently, recycled continuations are reset and initialized in freeContinuation at the end of their use. There are also gaps where the recycled continuation might retain old values and not be properly reset. The above approach leads to corruption and segfaults as seen in the runJavaThread failure reported in eclipse-openj9#16728. To fix the above crash, all continuations (new and recycled) are initialized and reset in createContinuation just before the continuation begins execution. Related: eclipse-openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
Depends on - eclipse-openj9/openj9#18649 - eclipse-openj9/openj9#18657 - eclipse-openj9/openj9#18658 Closes eclipse-openj9/openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
re #16728 (comment): Fix for ISSUE 1 and ISSUE 3
@LinHu2016 Can you confirm if issues 1 and 2 in #16728 (comment) are fixed by the above PRs? This issue can be closed after the below PRs are merged. Fix for ISSUE 2
Enable Skynet on AIX |
Depends on - eclipse-openj9/openj9#18649 - eclipse-openj9/openj9#18657 - eclipse-openj9/openj9#18658 Closes eclipse-openj9/openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
Depends on - eclipse-openj9/openj9#18649 - eclipse-openj9/openj9#18657 - eclipse-openj9/openj9#18658 Closes eclipse-openj9/openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
Depends on - eclipse-openj9/openj9#18649 - eclipse-openj9/openj9#18657 - eclipse-openj9/openj9#18658 Closes eclipse-openj9/openj9#16728 Signed-off-by: Babneet Singh <[email protected]>
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_sanity.openjdk_aarch64_linux_Nightly/111/
jdk_lang_0
java/lang/Thread/virtual/stress/Skynet.java
https://openj9-artifactory.osuosl.org/artifactory/ci-openj9/Test/Test_openjdk19_j9_sanity.openjdk_aarch64_linux_Nightly/111/openjdk_test_output.tar.gz
The text was updated successfully, but these errors were encountered: