[oe] [meta-java] openjre-8: java from master/master-next crashes on startup

[Jacob Kroon]

Hi,

For me, this has been the status for quite a while now. First of all, 
openjre-8 from master/master-next doesn't build on my Fedora 30 box, 
I've been using the patch attached to this mail to get it to build.
   Secondly, java always crashes on startup here:

> #0  Node::Node (this=0x805fa88, req=3)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/node.cpp:327
> #1  0xb7499b63 in RegionNode::RegionNode (required=3, this=0x805fa88)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/cfgnode.hpp:73
> #2  LoopNode::LoopNode (backedge=0x0, entry=0x0, this=0x805fa88)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/loopnode.hpp:88
> #3  RootNode::RootNode (this=0x805fa88)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/rootnode.hpp:37
> #4  Compile::Init (this=0x977feaf8, aliaslevel=0)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/compile.cpp:1079
> #5  0xb74a02c5 in Compile::Compile (this=0x977feaf8, ci_env=0x977ff188, 
>     generator=0xb7895270 <OptoRuntime::new_instance_Type()>, 
>     stub_function=0xb7897070 <OptoRuntime::new_instance_C(Klass*, JavaThread*)> "U\211\345WVS\350\245+\237\377\201\303\205\217\062", 
>     stub_name=0xb7a0a62f "_new_instance_Java", is_fancy_jump=0, pass_tls=true, 
>     save_arg_registers=false, return_pc=false)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/compile.cpp:1012
> #6  0xb789587b in OptoRuntime::generate_stub (env=0x977ff188, 
>     gen=0xb7895270 <OptoRuntime::new_instance_Type()>, 
>     C_function=0xb7897070 <OptoRuntime::new_instance_C(Klass*, JavaThread*)> "U\211\345WVS\350\245+\237\377\201\303\205\217\062", name=0xb7a0a62f "_new_instance_Java", 
>     is_fancy_jump=0, pass_tls=true, save_argument_registers=false, return_pc=false)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/runtime.cpp:181
> #7  0xb7895a5f in OptoRuntime::generate (env=0x977ff188)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/runtime.cpp:147
> #8  0xb7417de9 in C2Compiler::init_c2_runtime ()
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/runtime/thread.hpp:1854
> #9  0xb7417e35 in C2Compiler::initialize (this=0xb706b180)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/opto/c2compiler.cpp:102
> #10 0xb74a508c in CompileBroker::init_compiler_runtime ()
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/compiler/compileBroker.cpp:1674
> #11 0xb74ab6ed in CompileBroker::compiler_thread_loop ()
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/compiler/compileBroker.cpp:1772
> #12 0xb79238d5 in compiler_thread_entry (thread=0xb7079400, __the_thread__=0xb7079400)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/runtime/thread.cpp:3230
> #13 0xb792d782 in JavaThread::thread_main_inner (this=0xb7079400)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/share/vm/runtime/thread.hpp:1035
> #14 0xb7820ce4 in java_start (thread=0xb7079400)
>     at /usr/src/debug/openjre-8/172b11-r0/jdk8u-33d274a7dda0/hotspot/src/os/linux/vm/os_linux.cpp:790
> #15 0x4121244c in start_thread (arg=<optimized out>) at pthread_create.c:479
> #16 0x4113bd56 in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:108

The Node::Node() constructor uses an IDX_INIT() macro:

> #define IDX_INIT(req) this->Init((req), (Compile*) this->_out)

and Node::Init() does:

> 297     inline int Node::Init(int req, Compile* C) {
> 298	  assert(Compile::current() == C, "must use operator new(Compile*)");
> 299	  int idx = C->next_unique();

and here, this->_out = C = null, which I think leads to a segfault on 
line 299.
   I've tried looking at upstream jdk8u but can't find anything. Does 
anyone have an idea of whats going on here, or if I should raise the 
issue to openjdk ml ?

Cheers,
Jacob
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Fix-build.patch
Type: text/x-patch
Size: 2342 bytes
Desc: not available
URL: <http://lists.openembedded.org/pipermail/openembedded-devel/attachments/20191013/31372aea/attachment.bin>