JDK-8218468 : Load barrier slow path node should be MachTypeNode
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 11,11-shenandoah,13
  • Priority: P3
  • Status: Resolved
  • Resolution: Fixed
  • Submitted: 2019-02-05
  • Updated: 2020-01-01
  • Resolved: 2019-04-17
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 11 JDK 13
11.0.6-oracleFixed 13 b17Fixed
Related Reports
Duplicate :  
Duplicate :  
Sub Tasks
JDK-8219557 :  
Description
#  Internal Error (workspace/open/src/hotspot/share/opto/machnode.cpp:381), pid=4662, tid=4684
#  assert(tp->base() != Type::AnyPtr) failed: not a bare pointer
#
# JRE version: Java(TM) SE Runtime Environment (13.0) (fastdebug build 13-internal+0-jdk13-jdk.316)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (fastdebug 13-internal+0-jdk13-jdk.316, mixed mode, tiered, z gc, linux-amd64)
# Problematic frame:
# V  [libjvm.so+0x127ddb6]  MachNode::adr_type() const+0x646
#

Command Line: -XX:MaxRAMPercentage=6 -XX:+UnlockExperimentalVMOptions -XX:+UseZGC -XX:+IgnoreUnrecognizedVMOptions --add-opens=java.base/java.net=ALL-UNNAMED -Dseed=5532577772482251 -XX:MaxRAMPercentage=50 applications.runthese.Runner -duration 30 -runlist ./RunTheseTestList.dat

Current CompileTask:
C2:1762031 119083       4       java.lang.invoke.StringConcatFactory$MethodHandleInlineCopyStrategy::generate (497 bytes)

Stack: [0x00007f93ec66b000,0x00007f93ec76c000],  sp=0x00007f93ec766a60,  free space=1006k
Native frames: (J=compiled Java code, A=aot compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0x127ddb6]  MachNode::adr_type() const+0x646
V  [libjvm.so+0xd667da]  PhaseCFG::insert_anti_dependences(Block*, Node*, bool)+0x5a
V  [libjvm.so+0xd6e528]  PhaseCFG::schedule_late(VectorSet&, Node_Stack&)+0xb08
V  [libjvm.so+0xd6ebae]  PhaseCFG::global_code_motion()+0x38e
V  [libjvm.so+0xd6fa61]  PhaseCFG::do_global_code_motion()+0x51
V  [libjvm.so+0x9fe1ec]  Compile::Code_Gen()+0x24c
V  [libjvm.so+0xa01fc5]  Compile::Compile(ciEnv*, C2Compiler*, ciMethod*, int, bool, bool, bool, DirectiveSet*)+0xe65
V  [libjvm.so+0x80701e]  C2Compiler::compile_method(ciEnv*, ciMethod*, int, DirectiveSet*)+0xde
V  [libjvm.so+0xa0f5a2]  CompileBroker::invoke_compiler_on_method(CompileTask*)+0x402
V  [libjvm.so+0xa106d8]  CompileBroker::compiler_thread_loop()+0x458
V  [libjvm.so+0x17afc1f]  JavaThread::thread_main_inner()+0x2cf
V  [libjvm.so+0x17b8d8c]  JavaThread::run()+0x1cc
V  [libjvm.so+0x17b65c0]  Thread::call_run()+0x100
V  [libjvm.so+0x149eead]  thread_native_entry(Thread*)+0x10d

Comments
Fix Request (11u) This resolves compilation problem with ZGC enabled. Patch does not apply cleanly to 11u, but the adjustments are trivial. 11u RFR: https://mail.openjdk.java.net/pipermail/jdk-updates-dev/2019-September/001813.html
02-09-2019

Evaluating
14-03-2019

[~neliasso][~iignatyev][~jwilhelm] Just failed again. Please problem list this while the fix is being worked on.
21-02-2019

you either remove it from problemList in your workspace, or don't specify problemList in your run, or use problemList to select tests instead of excluding (CODETOOLS-7902326).
21-02-2019

[~iignatyev] I think it makes sense. How do you run a test that has been ProblemListed?
20-02-2019

I still haven't been able to reproduce locally, so my fix is just a qualified guess and I'm not happy with that.
20-02-2019

[~mikael] [~stefank] [~pliden], I have a generic question about zgc testing/stability, this is not the first case when we see something failing only w/ zgc and given zgc relatively young, this is to be expected. so I can't help but wonder if it makes sense to introduce a special problem list for zgc (in the same way we did for graal, noncds and Xcomp).
20-02-2019

[~neliasso] this seems to be failing frequently. Is the fix imminent, or should the test be problem listed/disabled meanwhile?
20-02-2019

http://cr.openjdk.java.net/~neliasso/8218468/webrev.01/
14-02-2019

I think I found the problem. Can't reproduce reliably however. The zLoadBarrierSlowRegNodes should be of MachTypeNode, not just MachNode.
14-02-2019