Loom fails tobuild on linux x64

Alan Bateman Alan.Bateman at oracle.com
Mon Jan 13 07:47:56 UTC 2020

On 13/01/2020 00:01, forax at univ-mlv.fr wrote:
> good idea !
> https://travis-ci.org/forax/java-next/jobs/636110527#L2141
That's useful as it has the backtrace:

2170V  [libjvm.so+0x35de96] Label::patch_instructions(MacroAssembler*)+0x116
2171V  [libjvm.so+0x101c522] 
StubGenerator::generate_disjoint_word_copy_up(bool, char const*)+0x962
2172V  [libjvm.so+0x1026212] 
2173V  [libjvm.so+0x103c6b8]  StubGenerator::generate_all()+0x10a8
2174V  [libjvm.so+0x1017130]  StubGenerator_generate(CodeBuffer*, int)+0x60
2175V  [libjvm.so+0x103ef16]  StubRoutines::initialize2() [clone 
2176V  [libjvm.so+0xb12e16]  init_globals()+0xf6
2177V  [libjvm.so+0x10c1db4]  Threads::create_vm(JavaVMInitArgs*, 
2178V  [libjvm.so+0xbc36d6]  JNI_CreateJavaVM+0x56
2179C  [libjli.so+0x47c1]  JavaMain+0x81
2180C  [libjli.so+0x8929]  ThreadJavaMain+0x9

StubGenerator::generate_disjoint_word_copy_up generates the code to copy 
a continuation from the stack so it doesn't exist in the main line. 
There are several places where it uses jmpb and jccb that might be the 
issue with some UseAVX values. I'm sure Ron or Rickard will recognize 
this quickly.


More information about the loom-dev mailing list