Fix redefinition failure on openj9 #5009
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
trask
reviewed
Jan 4, 2022
...main/java/io/opentelemetry/javaagent/instrumentation/internal/reflection/RealInterfaces.java
Outdated
Show resolved
Hide resolved
...main/java/io/opentelemetry/javaagent/instrumentation/internal/reflection/RealInterfaces.java
Outdated
Show resolved
Hide resolved
trask
reviewed
Jan 4, 2022
...main/java/io/opentelemetry/javaagent/instrumentation/internal/reflection/RealInterfaces.java
Outdated
Show resolved
Hide resolved
…whether it has any of the virutal field marker interfaces
...ava/io/opentelemetry/javaagent/instrumentation/internal/reflection/VirtualFieldDetector.java
Outdated
Show resolved
Hide resolved
mateuszrzeszutek
approved these changes
Jan 5, 2022
…entation is disabled
trask
approved these changes
Jan 5, 2022
| /** Helper class for detecting whether given class has virtual fields. */ | ||
| public final class VirtualFieldDetector { | ||
|
|
||
| private static final Cache<Class<?>, Boolean> classesWithVirtualFields = Cache.weak(); |
There was a problem hiding this comment.
As ClassValue is not mutable, it works best if you can compute the value on first use. Here we can't really compute the value, so the best we could do would be to use a mutable object as value of ClassValue, initialize it to false, and update it when the actual value is known. With current solution we can ignore the classes that don't have virtual fields, with ClassValue we would probably end up initializing a ClassValue for these too.
iNikem
approved these changes
Jan 6, 2022
RashmiRam
pushed a commit
to RashmiRam/opentelemetry-auto-instr-java
that referenced
this pull request
May 23, 2022
* Fix redefinition failure on openj9 * isntead of remembering the list of interfaces the class has remember whether it has any of the virutal field marker interfaces * address review comment * ensure virtual field detection works when internal-reflection instrumentation is disabled
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Resolves #4848 The redefinition failure in the linked case isn't what made the test fail, but as it seems to be impossible to tell what the actual cause is it is best to close it.
After hiding our interfaces from
Class.getInterfacesvirtual field implementation isn't able to tell that class has been transformed and thus skips transforming it which breaks redefinition as new bytes don't have the same interfaces as the redefined class.Although this stack trace has been only observed on openj9 it could happen on hotspot too.