KONNO Kazuhiro

I am not sure if this is the reason of the failure, but I see a garbage `^]` in the `Load command 13` of the `otool` output: ``` % otool...

I built x86-64 Linux runtimes to try this. Neither the compressed refs build nor the large heap build reproduced the assertion failure.

Yes, I am building debug builds locally. @0xdaryl says he was able to reproduce the failure on aarch64 with a release build by changing the `TR_ASSERT()` to `TR_ASSERT_FATAL()`. I am...

The assertion failure occurs during the loopVersioner optimization. It disappears by adding `disableLoopVersioner` to the `-Xjit:` option. ``` Performing 47: loopVersioner [ 417] O^O LOOP VERSIONER: Versioning natural loop 84...

Today I recreated the assertion failure on x86_64 Linux, and generated the attached trace files by running with the following command. The boundary of the subindex is between 422 and...

@fjeremic I just run `java -Xjit:optlevel=warm,count=0,limit=boot` specifying no class to run as I wrote in the original post. I get the message starting with "Usage: java [options] [args...]" when I...

Backtrace on aarch64 with debug information: ``` #4 TR::assertion ( file=file@entry=0x7fb6b53d78 "/root/openj9-openjdk-jdk11/build/linux-aarch64-normal-server-release/vm/compiler/../omr/compiler/il/OMRBlock.cpp", line=line@entry=277, condition=condition@entry=0x7fb6b540e0 "(frequency >= 0)", format=format@entry=0x7fb6b540b8 "Block frequency must be non negative\n") at /root/openj9-openjdk-jdk11/build/linux-aarch64-normal-server-release/vm/compiler/../omr/compiler/infra/Assert.cpp:157 #5 0x0000007fb66e1b90 in OMR::Block::createEmptyBlock...

I still can recreate the assertion failure with x86_64 Linux and aarch64 today. What can I do with this?

I still can recreate this assertion failure with a debug build on AArch64. (I just haven't tried x86-64 today).

I reproduced this assertion failure on x86-64 Linux using sanity.functional last month. See #17616.