[aarch64-port-dev ] JTREG hotspot issues

Andrew Haley aph at redhat.com
Tue Aug 27 10:20:31 PDT 2013


On 08/23/2013 08:17 PM, Andy Johnson wrote:
> The JTREG regression tests in hotspot/test are finding the following
> errors, all a result of finding Unimplemented stubs:
> 
> runtime/6294277/SourceDebugExtension.java
> #  Internal Error
> (~/openjdk-8/jdk8/hotspot/src/cpu/aarch64/vm/interp_masm_aarch64.cpp:54),
> pid=2981, tid=547984654864
> #  Error: Unimplemented()
> void InterpreterMacroAssembler::check_and_handle_popframe(Register
> java_thread) {
>   if (JvmtiExport::can_pop_frame()) {
>     Unimplemented();
>   }
> }
> 
> gc/8000311/Test8000311.java
> gc/TestG1ZeroPGCTJcmdThreadPrint
> gc/7168848/HumongousAlloc
> #  Internal Error
> (~/openjdk-8/jdk8/hotspot/src/cpu/aarch64/vm/macroAssembler_aarch64.cpp:2254),
> pid=1341, tid=548143497744
> #  Error: Unimplemented()
> void MacroAssembler::g1_write_barrier_pre(Register obj,
>                                           Register pre_val,
>                                           Register thread,
>                                           Register tmp,
>                                           bool tosca_live,
>                                           bool expand_call) {
> Unimplemented(); }
> 
> Is there any plan to implement these features?

The first one should be fixed already.  Please update your sources.

> There may be periodic
> failures in the G1 garbage collector if the latter is not implemented,
> unless there is a way to bypass this call, other than specifying an
> alternate garbage collector.
> 

We don't claim to support anything beyond Java's defaults right now.

However, if you can provide proper details of how to reproduce these
issues I can have a look.

Andrew.




More information about the aarch64-port-dev mailing list