Obtain can_generate_sampled_object_alloc_events JVMTI capability only when needed #1011
Conversation
This file contains hidden or 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
apangin
reviewed
Sep 26, 2024
… when needed Only one JVM TI agent can hold can_generate_sampled_object_alloc_events capability at a time. To avoid multiple profiler instances from conflicting with each other, profiler should obtain the capability only when needed and release it as soon as a profiling session finishes.
apangin
reviewed
Oct 1, 2024
test/test/alloc/AllocTests.java
Outdated
|
|
||
| @Test(mainClass = MapReaderOpt.class, jvmVer = {11, Integer.MAX_VALUE}) | ||
| public void objectSamplerWtihDifferentAsprofs(TestProcess p) throws Exception { | ||
| Thread.sleep(1000); |
There was a problem hiding this comment.
Why is that? If there is a problem with timing, it should be solved centrally, preferably in a more reliable way.
There was a problem hiding this comment.
I've removed the sleep
test/test/alloc/AllocTests.java
Outdated
| // _[k] suffix in collapsed output corresponds to jdk.ObjectAllocationOutsideTLAB, which means alloc tracer is being used | ||
| assert !out.contains("_\\[k\\]"); // we are using alloc tracer instead of object sampler, should definitely not happen on first profiling call | ||
| Path asprofCopy = Path.of("/tmp/libasyncProfiler." + p.currentOs().getLibExt()); | ||
| Files.copy(Path.of("build/lib/libasyncProfiler." + p.currentOs().getLibExt()), asprofCopy, StandardCopyOption.REPLACE_EXISTING); |
There was a problem hiding this comment.
It's better to encapsulate the logic for getting library path in TestProcess, especially since TestProcess already has it.
test/test/alloc/AllocTests.java
Outdated
| Output out = p.profile("-e alloc -d 3 -o collapsed"); | ||
| // _[k] suffix in collapsed output corresponds to jdk.ObjectAllocationOutsideTLAB, which means alloc tracer is being used | ||
| assert !out.contains("_\\[k\\]"); // we are using alloc tracer instead of object sampler, should definitely not happen on first profiling call | ||
| Path asprofCopy = Path.of("/tmp/libasyncProfiler." + p.currentOs().getLibExt()); |
There was a problem hiding this comment.
Temporary file should have a unique name (not to clash with an existing library and to be able to run multiple test processes simultaneously). The file should be deleted on exit.
apangin
reviewed
Oct 9, 2024
roy-soumadipta
pushed a commit
to roy-soumadipta/async-profiler
that referenced
this pull request
Oct 15, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Description
Only one JVM TI agent can hold can_generate_sampled_object_alloc_events capability at a time. To avoid multiple profiler instances from conflicting with each other, profiler should obtain the capability only when needed and release it as soon as a profiling session finishes.
Related issues
Motivation and context
This allows multiple different async profiler instances to use the object sampler on the same java process one after the other.
How has this been tested?
Verifying that the test fails without implementation changes: https://github.com/openorclose/async-profiler/actions/runs/11050192717/job/30697437468#step:6:51
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.