JDK-7147744 : CTW: assert(false) failed: infinite EA connection graph build
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: hs23
  • Priority: P3
  • Status: Closed
  • Resolution: Fixed
  • OS: generic
  • CPU: generic
  • Submitted: 2012-02-22
  • Updated: 2013-07-18
  • Resolved: 2012-03-24
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 7 JDK 8 Other
7u40Fixed 8Fixed hs24Fixed
Related Reports
Duplicate :  
Relates :  
Relates :  
Relates :  
Description
CTW fails with crash. Reproduces starting HS 23 b16 (not with JDK 8 b26, HS 23 b15):

# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (/tmp/jprt/P1/230908.jcoomes/source/src/share/vm/opto/escape.cpp:1720), pid=27559, tid=3043085200
#  assert(false) failed: infinite EA connection graph build (86.155175 sec, 3 iterations) with 23111 nodes and worklist size 2211
#
# JRE version: 8.0-b26
# Java VM: Java HotSpot(TM) Server VM (23.0-b16-internal-201202172309.jcoomes.hs23-b16-snapshot-fastdebug mixed mode linux-x86 )
# Core dump written. Default location: /export/local/58270.HSX.PIT.VM_linux-i586_ctw__server_comp_ibiblio_6800_6900.testlist/results/ctw_maven2_com_yahoo_platform_yui_yuicompressor_2_3_6_yuicompressor-2_3_6_jar_classes_1_313/core or core.27559

Current thread (0x0826e000):  JavaThread "C2 CompilerThread0" daemon [_thread_in_native, id=27577, stack(0xb559c000,0xb561d000)]

Stack: [0xb559c000,0xb561d000],  sp=0xb561a2a0,  free space=504k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0xb197b1]  VMError::report_and_die()+0x1a1;;  VMError::report_and_die()+0x1a1
V  [libjvm.so+0x4f0fd8]  report_vm_error(char const*, int, char const*, char const*)+0x68;;  report_vm_error(char const*, int, char const*, char const*)+0x68
V  [libjvm.so+0x57d7a4]  ConnectionGraph::compute_escape()+0xcc4;;  ConnectionGraph::compute_escape()+0xcc4
V  [libjvm.so+0x57f437]  ConnectionGraph::do_analysis(Compile*, PhaseIterGVN*)+0xb7;;  ConnectionGraph::do_analysis(Compile*, PhaseIterGVN*)+0xb7
V  [libjvm.so+0x47bdd3]  Compile::Optimize()+0xb63;;  Compile::Optimize()+0xb63
V  [libjvm.so+0x47fe3c]  Compile::Compile(ciEnv*, C2Compiler*, ciMethod*, int, bool, bool)+0x14bc;;  Compile::Compile(ciEnv*, C2Compiler*, ciMethod*, int, bool, bool)+0x14bc
V  [libjvm.so+0x392729]  C2Compiler::compile_method(ciEnv*, ciMethod*, int)+0x229;;  C2Compiler::compile_method(ciEnv*, ciMethod*, int)+0x229
V  [libjvm.so+0x488888]  CompileBroker::invoke_compiler_on_method(CompileTask*)+0x598;;  CompileBroker::invoke_compiler_on_method(CompileTask*)+0x598
V  [libjvm.so+0x48984b]  CompileBroker::compiler_thread_loop()+0x5bb;;  CompileBroker::compiler_thread_loop()+0x5bb
V  [libjvm.so+0xaafa2f]  compiler_thread_entry(JavaThread*, Thread*)+0x4f;;  compiler_thread_entry(JavaThread*, Thread*)+0x4f
V  [libjvm.so+0xabbbb9]  JavaThread::thread_main_inner()+0x1b9;;  JavaThread::thread_main_inner()+0x1b9
V  [libjvm.so+0xabc24e]  JavaThread::run()+0x26e;;  JavaThread::run()+0x26e
V  [libjvm.so+0x939871]  java_start(Thread*)+0x111;;  java_start(Thread*)+0x111
C  [libpthread.so.0+0x61b5]  start_thread+0xc5


Current CompileTask:
C2: 188792 1452   !b  4       org.mozilla.javascript.Interpreter::interpretLoop (7845 bytes)

Comments
EVALUATION http://hg.openjdk.java.net/hsx/hotspot-rt/hotspot/rev/ee138854b3a6
28-03-2012

EVALUATION I rewrote Connection graph construction code in EA to reduce time spent there. In the bug's test time reduced by 100 (from about 50 to .5 sec on Nahalem-EX machine). Connection graph now has specialized classes for nodes and additional use edges to put on worklist only uses of node which added new point edge. Field node has also bases edges. Edges never removed only added. Instead of looking for Field's bases from the start create simple base edge to LocalVar during initial graph construction in build_connection_graph(). Late do several iteration to push all known JavaObject nodes references through graph. This phase has limits on number and time. Also on each iteration check if there are non globally escaped objects and bail out from code if not. Added additional Arraycopy node to connect source and destination objects. I removed uncast() calls so that all LocalVar nodes point to all related JavaObject nodes. I combined record_for_escape_analysis() and build_connection_graph() into one method. Added TracePhase around Connection graph build code to see how much time spent there. This code need addition work since I still saw outlier (10 min in EA) in sje2010 on SPARC. But I will look on it after this one is done. I added new GrowableArray method delete_at(i) to avoid shifting following elements which is done in remove_at(i) method. And changed GrowableArray::append_if_missing() to return true if a value was missed and was added into array.
12-03-2012

EVALUATION http://hg.openjdk.java.net/hsx/hotspot-comp/hotspot/rev/ee138854b3a6
12-03-2012

EVALUATION With new implementaion of EA Connection construction I am working on the time reduced from 44 sec to 0.3 sec: 8027 9 !b org.mozilla.javascript.Interpreter::interpretLoop (7845 bytes) 12958 9 size: 243136 time: 4929 inlined: 135 bytes <phase name='escapeAnalysis' nodes='24803' stamp='8.311'> <phase name='buildConnectionGraph' nodes='24803' stamp='8.312'> <phase_done nodes='24803' stamp='8.613'/> </phase> <phase_done nodes='24803' stamp='8.613'/> </phase>
23-02-2012