Deadlock prevention on @NotCheckpointSafe frames - #15540
Conversation
b913b55 to
900fff1
Compare
|
@gacholio Please take a look at these changes |
|
Jenkins test sanity xlinuxcriu jdk11 |
|
You should probably update the DDR rom dumper with the new bit. |
I couldnt find an example of this for other annotations like |
|
It's entirely possible that they were missed too. I believe the DDR dumper is used far far more than cfdump these days, so that's where we really need the proper decode. |
I created a new PR to address that, #15561 |
ae7d0a6 to
f2246d1
Compare
|
You can launch appropriate testing. |
|
Jenkins test sanity win jdk8 |
|
Jenkins test sanity xlinuxcriu jdk11 |
|
|
Jenkins test sanity win jdk8 |
|
link to the xlinux build https://openj9-jenkins.osuosl.org/job/PullRequest-OpenJ9/2439/ |
There are certain operations in the JCL that acquire a global lock (ie. methodType interning). The PR introduces a new mechanism to delay taking a checkpoint (for a certain amount of attempts) while a non checkpoint thread owns the global resource. Fixes: eclipse-openj9#15367 Signed-off-by: Tobi Ajila <atobia@ca.ibm.com>
|
Jenkins test sanity win jdk8 |
|
The xlinux build passed as well |
Deadlock prevention on @NotCheckpointSafe frames
There are certain operations in the JCL that acquire a global lock (ie.
methodType interning). The PR introduces a new mechanism to delay taking
a checkpoint (for a certain amount of attempts) while a non checkpoint
thread owns the global resource.
Fixes: #15367
Signed-off-by: Tobi Ajila atobia@ca.ibm.com