-
Notifications
You must be signed in to change notification settings - Fork 721
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
Enable native thread collection in javacores on OSX #14201
base: master
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This PR depends on eclipse/omr#6267. The code changes LGTM.
@EricYangIBM This PR was closed due to the deletion of |
Looks like I accidentally deleted my branches when trying to push a renamed branch |
cf29435
to
ff56a38
Compare
Is |
As per cmake/modules/platform/arch/aarch64.cmake#L23-L24, both |
e6a824b
to
7666837
Compare
e609f71
to
82af0bd
Compare
82af0bd
to
34bdd64
Compare
The related OMR changes have been merged: |
@tajila Can you help with merging this PR? |
@EricYangIBM Is this change to enable native thread collection on x86 mac osx? |
Yes, it enables the omrintrospect setup for the thread collection. The |
Okay, can you please add more details to the commit message + description describing this. I wasn't sure why this change was being made |
And just to be clear, support on aarch64 mac osx is still pending? |
Yes |
Enable native thread/stack iteration on x86 OSX, i.e. the omrintrospect functions used to collect backtraces in the javacore thread section. Omrintrospect library for OSX is implemented in eclipse/omr#6267 Signed-off-by: Eric Yang <eric.yang@ibm.com>
34bdd64
to
5eea6e9
Compare
Jenkins test sanity osx jdk8 |
Five different timeouts occurred while invoking
|
@babsingh are these failures expected? |
No, they are not expected. These timeouts were not seen before merging the OMR changes: eclipse/omr#6267 (comment) and eclipse/omr#6467 (comment). These failures are new, and they happen during javacore generation. |
I can't reproduce the failures. |
Next step: request access to the machine and verify. ++@AdamBrousseau for accessing the machines involved in https://openj9-jenkins.osuosl.org/job/Test_openjdk8_j9_sanity.functional_x86-64_mac_Personal/115/. |
I found that it is hanging at J9JavaVM->walkStackFrames on the second last thread openj9/runtime/rasdump/javadump.cpp Line 4430 in 9b82fcc
openj9/runtime/rasdump/javadump.cpp Line 5647 in 9b82fcc
|
|
I can't get any useful frames
I'm using |
This work can paused for the time being. Cgroup v2 takes priority and needs to be completed by the end of May. @EricYangIBM Focus on cgroup v2. I will take a look at the outlier failure/hang by the end of this week. |
@gacholio Would you have insights about this failure (happens during Finally got to look at the hang.
Last output in javacore at segfault:
lldb output at segfault:
|
No real insight - what you're seeing would seem to be impossible. Can you dump the walkState at the crashing point? I verified that the code uses walkThread (must always be valid) rather than currentThread (sometimes NULL during javacode production) to fetch the port library. This smacks of corruption, though why it would only show up on OSX I can't say. Looking into the walkState details will tell us more. |
@gacholio In the below
WalkState during segfault
|
Enable native thread/stack iteration on x86 OSX, i.e. the omrintrospect
functions used to collect backtraces in the javacore thread section.
Omrintrospect library for OSX is implemented in eclipse/omr#6267
Signed-off-by: Eric Yang eric.yang@ibm.com