-
Notifications
You must be signed in to change notification settings - Fork 733
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
j9mm.479 ASSERTION FAILED at gc_glue_java/ScavengerRootScanner.hpp:109: ((MM_StackSlotValidator(MM_StackSlotValidator::NOT_ON_HEAP, *slotPtr, stackLocation, walkState).validate(_env))) #16510
Comments
There are a few bad slots in the same thread stack, all of them have the same value:
|
@0xdaryl FYI |
JDK17 internal build(
|
this problem looks similar we observed earlier:
Interestedly, there is inlined object at
|
I started a 20x Grinder job on AArch64 Linux (internal job/Grinder/30827), but it seems it takes nearly an hour for running jdk_vector_1. The job might be killed by timeout after 20 hours. |
2 failures out of 18 iterations in the Grinder job above. It is unclear at this point if the failures are related to vector implementation or not. |
I ran Int64VectorTests.java on AArch64 Linux in my environment, using the binary from #16510 (comment). |
From the jdmpview "info jitm" output:
There are two versions of JIT generated code for this method. |
I ran jdk_vector_1 on AArch64 Linux in my environment 20 times, and got no test failures. |
I ran 3x20 Grinder jobs on AArch64 macOS (31570, 31571, 31573) with the following option to take the JIT trace file, but I was not able to reproduce the failure.
|
Another Grinder job (31579) fails with a different Java method,
From the jdmpview "info jitm" output:
The size of the first one is 85KB, and the second one is only 584 bytes long. "PC=000000013034ACAC" is in the range of the first one.
|
Another failure in the job 31579 with Vector128ConversionTests:
The output above points to the method
In both cases, "O-Slot value" points to an address in the stack near the "O-Slot" value, but the distance ("O-Slot" - "O-Slot value") is different. |
I don't think it guarantees anything, but I got no failures in 3x20 Grinder jobs with the |
2 more x20 Grinder jobs with |
As this is an intermittent failure only seen in an incubator feature this should not hold up 0.38. Moving out to 0.40. |
JDK17 aarch64_mac 0.38 m2 build(
|
JDK19 aarch64_linux(
|
Again, the wrong object reference (O-Slot value=0x225524) points to somewhere between O-Slot (0x2255C0) and walkSP (0x225328). The call stack of the MainThread looks like this:
The size of the compiled method
|
There is another manifestation: #17601 |
|
The failure with |
PR #17825 has been merged. |
It is almost 2 weeks since PR #17825 was merged. I think this issue can be closed now. |
Not seen in recent builds. |
JDK8 s390x_linux 0.41 milestone 1(
|
@JasonFengJ9 I think it is better to open a separate issue for the failure on zLinux above. |
Yeah, this issue occurred only on |
JDK17 internal build(
macaarch64rt8
)300x grinder - the assertion failure was reproduced while there were infra errors
Originally posted by @JasonFengJ9 in #15251 (comment)
The text was updated successfully, but these errors were encountered: