JDK-8029443 : 'assert(klass->is_loader_alive(_is_alive)) failed: must be alive' during VM_CollectForMetadataAllocation
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: hs25,8u40,9
  • Priority: P3
  • Status: Resolved
  • Resolution: Fixed
  • CPU: sparc
  • Submitted: 2013-12-03
  • Updated: 2015-01-21
  • Resolved: 2014-08-06
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 8 JDK 9
8u40Fixed 9 b28Fixed
Related Reports
Relates :  
JDK-8055657 - Test compiler/classUnloading/methodUnloading/TestMethodUnloading.java does not work with non-default GC
Relates :  
JDK-8054402 - "klass->is_loader_alive(_is_alive)) failed: must be alive" for anonymous classes
Description
runThese triggered assert
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (/tmp/jprt/P1/104812.anoll/s/src/share/vm/code/nmethod.cpp:1737), pid=13805, tid=46
#  assert(klass->is_loader_alive(_is_alive)) failed: must be alive
#
# JRE version: Java(TM) SE Runtime Environment (8.0-b118) (build 1.8.0-ea-fastdebug-b118)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.0-b62-internal-201312021048.anoll.8028109-fastdebug mixed mode solaris-sparc )
# Core dump written. Default location: /bpool/export/local/aurora/sandbox/results/runThese/core or core.13805
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.sun.com/bugreport/crash.jsp
#

---------------  T H R E A D  ---------------

Current thread (0x0000000100455800):  VMThread [stack: 0xffffffff43800000,0xffffffff43900000] [id=46]

Stack: [0xffffffff43800000,0xffffffff43900000],  sp=0xffffffff438fe080,  free space=1016k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0x123f1dc]  void VMError::report_and_die()+0x714
V  [libjvm.so+0x71fa28]  void report_vm_error(const char*,int,const char*,const char*)+0x78
V  [libjvm.so+0xebc504]  void CheckClass::check_class(Metadata*)+0x114
V  [libjvm.so+0xeb4dc4]  void nmethod::metadata_do(void(*)(Metadata*))+0x5ec
V  [libjvm.so+0xeb47b4]  void nmethod::verify_metadata_loaders(unsigned char*,BoolObjectClosure*)+0x8f4
V  [libjvm.so+0xeb3e9c]  void nmethod::do_unloading(BoolObjectClosure*,bool)+0xd8c
V  [libjvm.so+0x63f714]  void CodeCache::do_unloading(BoolObjectClosure*,bool)+0x7c
V  [libjvm.so+0xfe2cf0]  void PSParallelCompact::marking_phase(ParCompactionManager*,bool,ParallelOldTracer*)+0x5d0
V  [libjvm.so+0xfe1608]  bool PSParallelCompact::invoke_no_policy(bool)+0x670
V  [libjvm.so+0xfe0f30]  void PSParallelCompact::invoke(bool)+0x158
V  [libjvm.so+0x646c04]  void CollectedHeap::collect_as_vm_thread(GCCause::Cause)+0x1ac
V  [libjvm.so+0x124257c]  void VM_CollectForMetadataAllocation::doit()+0x224
V  [libjvm.so+0x126f378]  void VM_Operation::evaluate()+0xf8
V  [libjvm.so+0x126b2bc]  void VMThread::evaluate_operation(VM_Operation*)+0x254
V  [libjvm.so+0x126bd7c]  void VMThread::loop()+0x6a4
V  [libjvm.so+0x126ad24]  void VMThread::run()+0xe4
V  [libjvm.so+0xf1d308]  java_start+0x258

VM_Operation (0xffffffff2d8f9c28): CollectForMetadataAllocation, mode: safepoint, requested by thread 0x00000001078fb800
Comments
I figured out why this is only happening with -XX:-CompressedOops on Sparc: - With compressed oops the Klass is loaded by a 'loadConNKlass' instruction that calls MacroAssembler::set_narrow_klass(..) which then creates a metadata_Relocation. This metadata_Relocation is processed by CodeBuffer::finalize_oop_references(..) and an oop to the Klass is added. This oop causes the unloading of the method. - Without compressed oops we load the Klass using 'loadConP_no_oop_cheap' instead of 'loadConP_load' because isa_oop_ptr() is false. This instruction does not add a metadata relocation and therefore we do not add an oop in CodeBuffer::finalize_oop_references(..). Without an oop the nmethod is not unloaded, leading to stale metadata as described in my last comment. -> This is clearly a bug in the .ad file. Adding the appropriate relocation to 'loadConP_no_oop_cheap' fixes the bug. The bug also shows up on x86 with a 32 bit client version of the VM. I'm able to (rarely) reproduce it and suspect that this is a different bug in the C1 compiler.
28-07-2014
The problem is that we have stale metadata in a to-interpreter-stub that is used by an optimized virtual call (opt_virtual_call_type). In nmethod::do_unloading(..) we clean inline caches that refer to an unloaded Klass or Method but we don't do this for the optimized case. I don't understand why. Afterwards, we invoke nmethod::verify_metadata_loaders(..) to verify that there are no references to stale metadata left. We hit an assert because the Method referenced by the to-interpreter-stub belongs to an unloaded class. Note that we clean the stubs before verification to trim out dangling references to Method's, but only in the case the corresponding call is not in to-interpreted mode (see nmethod.cpp line 1739). Because otherwise the stub may still be called.
11-07-2014
I'm able to reproduce the bug with build 2014-07-03-003240.sspitsyn.rt-merge-to-hs on Solaris (Sparc v9). The nmethod java.lang.reflect.Constructor.newInstance is processed by CodeCache::do_unloading(..). Because it is not unloaded, nmethod::verify_metadata_loaders(..) checks that all of its metadata is still alive. The metadata section contains the Method "newInstance" of Klass sun/reflect/GeneratedConstructorAccessor3. Because the corresponding class loader is not alive, the assert is hit.
08-07-2014
The rebuilt fastdebug binaries do work. Debugging the client crash on sc11136750 nmethod for java/util/stream/StreamSpliterators$ArrayBuffer$OfInt.forEach has an unloaded (mirror) klass in the inline cache CompiledIC. It's not an oop, these should be renamed! 1823 Metadata* ic_oop = ic->cached_metadata(); 1824 if (ic_oop != NULL) { 1825 f(ic_oop); (gdb) x 0xf71f4488 (ic_oop) 0xf71f4488 <_ZTV19InstanceMirrorKlass+328>: 0xf686e330 I don't know that much about inline caches but if we put metadata in the CompiledICHolder we're should also have an oop in the oop section. I really want to give this back to the compiler group to check out this method (compiled by c1).
09-01-2014
Please take a look at this rebuild from the source archive. If the problem can not be reproduced, close this report as such. I'm sorry we don't have the storage space to keep everything in JPRT - I wish we did. I continue to bring this issue up with my management. Please do the same with your managers, and hopefully it will help us justify adding resources. -------- Original Message -------- Subject: JPRT: [hotspotwest] job notification - success with job 2014-01-07-193810.tbell.2013-11-22-220327-amurillo-hs25-b61-comp-sync Date: Tue, 7 Jan 2014 12:24:54 -0800 (PST) From: jprt_admin_ww_grp@oracle.com To: alejandro.murillo@oracle.com, coleen.phillimore@oracle.com, ioi.lam@oracle.com CC: jprt_notify_ww_grp@oracle.com JPRT: [hotspotwest] job notification - success with job 2014-01-07-193810.tbell.2013-11-22-220327-amurillo-hs25-b61-comp-sync JPRT Job ID: 2014-01-07-193810.tbell.2013-11-22-220327-amurillo-hs25-b61-comp-sync JPRT System: hotspotwest JPRT Version: 3.3.88: (2013-12-19) Case of the Blonde Bonanza [f6683e483a3b] Job URL: http://prt-web.us.oracle.com//archive/2014/01/2014-01-07-193810.tbell.2013-11-22-220327-amurillo-hs25-b61-comp-sync Job ARCHIVE: /net/prt-archiver.us.oracle.com/data/jprt/archive/2014/01/2014-01-07-193810.tbell.2013-11-22-220327-amurillo-hs25-b61-comp-sync User: tbell - alejandro.murillo@oracle.com ioi.lam@oracle.com coleen.phillimore@oracle.com Release: jdk8 Source: Source Tree: /var/tmp/2013-11-22-220327.amurillo.hs25-b61-comp-sync/{.} File List: {.} Command Line: jprt submit -noquickabort -precious -c 'rebuild of job 2013-11-22-220327.amurillo.hs25-b61-comp-sync to create fastdebug bits' -email alejandro.murillo@oracle.com -stree . -nocrupdates -extratime 8m -onlytests '.*scimark.*' -email ioi.lam@oracle.com -email coleen.phillimore@oracle.com Flags: PRECIOUS NORMAL Submitted at: Tuesday January 7, 2014 11:38:11 PST Time in queue: 46m 43s Started at: Tuesday January 7, 2014 11:38:52 PST Finished at: Tuesday January 7, 2014 12:24:54 PST Run time: 46m 01s Job state: success Flags: PRECIOUS NORMAL Bundles: USE: jprt install 2014-01-07-193810.tbell.2013-11-22-220327-amurillo-hs25-b61-comp-sync Run time: 46m 01s Build Stats: 32 pass, 0 fail, 0 killed, 0 working, 0 initializing, 0 not started Test Stats: 18 pass, 0 fail, 0 killed, 0 working, 0 initializing, 0 not started User Comments: rebuild of job 2013-11-22-220327.amurillo.hs25-b61-comp-sync to create fastdebug bits Build (Release:jdk8 Boot:<jdk7u7> Import:<jdk8>): linux_armsflt_2.6-fastdebugEmb success(12m 37s) USED: hostname=sc11137353 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=2s init=3s work=12m26s fini=5s NEEDS: ant171,armxcompiler,ccache,cups,ejdk8linuxarmsflt,freetype32,gnumake381,jmc8,npapi linux_armsflt_2.6-productEmb success(18m 17s) USED: hostname=sc11136051 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=2 parallelcount=2 ram=3042MB instance=P1 TIMING: clean=1s init=14s work=17m55s fini=7s NEEDS: ant171,armxcompiler,ccache,cups,ejdk8linuxarmsflt,freetype32,gnumake381,jmc8,npapi linux_armvfphflt_2.6-fastdebugEmb success(22m 41s) USED: hostname=sc11137352 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=2s init=23s work=22m8s fini=7s NEEDS: ant171,armxcompiler,ccache,ejdk8linuxarmvfphflt,freetype32,gnumake381,jmc8,npapi linux_armvfphflt_2.6-productEmb success(16m 25s) USED: hostname=sca00aez platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=3s work=16m10s fini=9s NEEDS: ant171,armxcompiler,ccache,ejdk8linuxarmvfphflt,freetype32,gnumake381,jmc8,npapi linux_armvfpsflt_2.6-fastdebugEmb success(20m 57s) USED: hostname=sc11137354 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=2s work=20m45s fini=7s NEEDS: ant171,armxcompiler,ccache,ejdk8linuxarmvfpsflt,freetype32,gnumake381,jmc8,npapi linux_armvfpsflt_2.6-productEmb success(19m 04s) USED: hostname=sc11137353 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=53s work=17m58s fini=9s NEEDS: ant171,armxcompiler,ccache,ejdk8linuxarmvfpsflt,freetype32,gnumake381,jmc8,npapi linux_armvh_2.6-fastdebug success(17m 38s) USED: hostname=sca00afa platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=3s init=22s work=17m4s fini=8s NEEDS: ant171,armxcompiler,ccache,freetype32,gnumake381,jdk8linuxarmvfphflt,jmc8,npapi linux_armvh_2.6-product success(16m 03s) USED: hostname=sc11137354 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=3s work=15m49s fini=8s NEEDS: ant171,armxcompiler,ccache,freetype32,gnumake381,jdk8linuxarmvfphflt,jmc8,npapi linux_i586_2.6-fastdebug success(21m 11s) USED: hostname=sc11136052 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=2 parallelcount=2 ram=3042MB instance=P1 TIMING: clean=1s init=41s work=20m22s fini=5s NEEDS: ant171,ccache,cups,freetype32,fxsdk8lin32,gnumake381,jmc8,npapi linux_i586_2.6-fastdebugEmb success(11m 36s) USED: hostname=sc11137352 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=3s init=3s work=11m24s fini=5s NEEDS: ant171,ccache,cups,freetype32,fxsdk8lin32,gnumake381,jmc8,npapi linux_i586_2.6-product success(08m 39s) USED: hostname=sca00afa platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=30s work=7m54s fini=11s NEEDS: ant171,ccache,cups,freetype32,fxsdk8lin32,gnumake381,jmc8,npapi linux_i586_2.6-productEmb success(10m 55s) USED: hostname=sc11137352 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=48s work=9m52s fini=12s NEEDS: ant171,ccache,cups,freetype32,fxsdk8lin32,gnumake381,jmc8,npapi linux_ppc_2.6-fastdebugEmb success(05m 45s) USED: hostname=sc11137354 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=3s work=5m34s fini=5s NEEDS: ant171,ccache,cups,ejdk8linuxppc,freetype32,gnumake381,jmc8,npapi,ppcxcompiler linux_ppc_2.6-productEmb success(08m 24s) USED: hostname=sc11136051 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=2 parallelcount=2 ram=3042MB instance=P1 TIMING: clean=1s init=14s work=8m1s fini=7s NEEDS: ant171,ccache,cups,ejdk8linuxppc,freetype32,gnumake381,jmc8,npapi,ppcxcompiler linux_ppcsflt_2.6-fastdebugEmb success(05m 00s) USED: hostname=sca00aez platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=2s init=2s work=4m50s fini=5s NEEDS: ant171,ccache,cups,ejdk8linuxppcsflt,freetype32,gnumake381,jmc8,npapi,ppcxcompiler linux_ppcsflt_2.6-productEmb success(08m 28s) USED: hostname=sc11136052 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=2 parallelcount=2 ram=3042MB instance=P1 TIMING: clean=1s init=13s work=8m5s fini=8s NEEDS: ant171,ccache,cups,ejdk8linuxppcsflt,freetype32,gnumake381,jmc8,npapi,ppcxcompiler linux_ppcv2_2.6-fastdebugEmb success(06m 16s) USED: hostname=sc11137353 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=3s init=2s work=6m5s fini=5s NEEDS: ant171,ccache,cups,ejdk8linuxppcv2,freetype32,gnumake381,jmc8,npapi,ppcxcompiler linux_ppcv2_2.6-productEmb success(04m 42s) USED: hostname=sca00aez platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=1s init=10s work=4m24s fini=5s NEEDS: ant171,ccache,cups,ejdk8linuxppcv2,freetype32,gnumake381,jmc8,npapi,ppcxcompiler linux_x64_2.6-fastdebug success(05m 24s) USED: hostname=sca00afg platform=linux_x64_2.6 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=16089MB instance=P1 TIMING: clean=2s init=2s work=5m16s fini=4s NEEDS: ant171,ccache,cups,freetype64,fxsdk8lin64,gnumake381,jmc8,npapi linux_x64_2.6-product success(05m 28s) USED: hostname=sca00aff platform=linux_x64_2.6 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=16089MB instance=P1 TIMING: clean=2s init=3s work=5m10s fini=11s NEEDS: ant171,ccache,cups,freetype64,fxsdk8lin64,gnumake381,jmc8,npapi linux_x64_2.6-productOpen success(05m 12s) USED: hostname=sca00afg platform=linux_x64_2.6 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=16089MB instance=P1 TIMING: clean=2s init=3s work=4m57s fini=9s NEEDS: ant171,ccache,cups,freetype64,fxsdk8lin64,gnumake381,jmc8,npapi macosx_x64_10.7-fastdebug success(07m 02s) USED: hostname=sc21jprtmac10 platform=macosx_x64_10.7 osname=macosx osarch=x64 cpus=8 parallelcount=8 instance=P1 TIMING: clean=1s init=7s work=6m44s fini=8s NEEDS: ant171,cups,fxsdk8mac64,gnumake381,jmc8,npapi,signing,sparkle macosx_x64_10.7-product success(07m 16s) USED: hostname=sc21jprtmac05 platform=macosx_x64_10.8 osname=macosx osarch=x64 cpus=4 parallelcount=4 instance=P1 TIMING: clean=1s init=7s work=6m57s fini=9s NEEDS: ant171,cups,fxsdk8mac64,gnumake381,jmc8,npapi,signing,sparkle solaris_sparcv9_5.10-fastdebug success(16m 40s) USED: hostname=sca00adn platform=solaris_sparc_5.10 osname=solaris osarch=sparc cpus=6 parallelcount=6 ram=31900MB instance=P1 TIMING: clean=4s init=21s work=16m7s fini=6s NEEDS: SS121,ant171,ccache,cups,freetype64,gnumake381,jmc8,npapi solaris_sparcv9_5.10-product success(18m 18s) USED: hostname=sc11136049 platform=solaris_sparc_5.10 osname=solaris osarch=sparc cpus=128 parallelcount=32 ram=65312MB instance=P1 TIMING: clean=3s init=7s work=17m59s fini=8s NEEDS: SS121,ant171,ccache,cups,freetype64,gnumake381,jmc8,npapi solaris_x64_5.10-debugOpen success(08m 40s) USED: hostname=sc11b2209 platform=solaris_i586_5.10 osname=solaris osarch=i586 cpus=4 parallelcount=4 ram=24567MB instance=P1 TIMING: clean=1s init=4s work=8m25s fini=8s NEEDS: SS121,ant171,ccache,cups,freetype64,gnumake381,jmc8,npapi solaris_x64_5.10-fastdebug success(10m 47s) USED: hostname=prt-host-x64-2 platform=solaris_i586_5.10 osname=solaris osarch=i586 cpus=15 parallelcount=15 ram=32192MB instance=P1 TIMING: clean=1s init=3s work=10m33s fini=7s NEEDS: SS121,ant171,ccache,cups,freetype64,gnumake381,jmc8,npapi solaris_x64_5.10-product success(12m 59s) USED: hostname=sc11b2210 platform=solaris_i586_5.10 osname=solaris osarch=i586 cpus=4 parallelcount=4 ram=24567MB instance=P1 TIMING: clean=1s init=4s work=12m46s fini=5s NEEDS: SS121,ant171,ccache,cups,freetype64,gnumake381,jmc8,npapi windows_i586_6.1-fastdebug success(11m 52s) USED: hostname=sc11136418 platform=windows_x64_6.1 osname=windows osarch=x64 cpus=2 parallelcount=2 ram=7799MB instance=P1 compiler=VS2010 cygwin=true pstools=true ATTRS: compiler=VS2010 TIMING: clean=1m42s init=3m2s work=7m4s fini=3s NEEDS: ant171,cygwin,freetype32,fxsdk8win32,gnumake381,jmc8,lzma,npapi,signing,sponsor_binaries1,sponsor_binaries2,upx,wix35 windows_i586_6.1-product success(07m 07s) USED: hostname=sc11136526 platform=windows_x64_6.1 osname=windows osarch=x64 cpus=2 parallelcount=2 ram=7799MB instance=P1 compiler=VS2010 mks=true cygwin=true installshield=true dxsdk=true pstools=true ATTRS: compiler=VS2010 TIMING: clean=7s init=57s work=5m55s fini=7s NEEDS: ant171,cygwin,freetype32,fxsdk8win32,gnumake381,jmc8,lzma,npapi,signing,sponsor_binaries1,sponsor_binaries2,upx,wix35 windows_x64_6.1-fastdebug success(09m 21s) USED: hostname=sca00aoa platform=windows_x64_6.1 osname=windows osarch=x64 cpus=6 parallelcount=6 ram=29999MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 cygwin=true pstools=true ATTRS: compiler=VS2010 TIMING: clean=1m43s init=2m45s work=4m48s fini=3s NEEDS: ant171,cygwin,freetype64,fxsdk8win64,gnumake381,jmc8,npapi,sponsor_binaries1,sponsor_binaries2,wix35 windows_x64_6.1-product success(08m 06s) USED: hostname=sca00aog platform=windows_x64_6.1 osname=windows osarch=x64 cpus=6 parallelcount=6 ram=29999MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 cygwin=true pstools=true ATTRS: compiler=VS2010 TIMING: clean=38s init=2m48s work=4m34s fini=5s NEEDS: ant171,cygwin,freetype64,fxsdk8win64,gnumake381,jmc8,npapi,sponsor_binaries1,sponsor_binaries2,wix35 Tests: linux_i586-fastdebug-c1-scimark success( 39s) USED: hostname=sca00afa platform=linux_i586_2.6 osname=linux osarch=i586 cpus=4 parallelcount=4 ram=3034MB instance=P1 TIMING: clean=2s init=6s work=30s fini=1s VMFLAGS: -client NEEDS: gnumake381,jtreg,scimark linux_i586-fastdebug-c2-scimark success( 40s) USED: hostname=sc11136051 platform=linux_i586_2.6 osname=linux osarch=i586 cpus=2 parallelcount=2 ram=3042MB instance=P1 TIMING: clean=1s init=7s work=28s fini=1s VMFLAGS: -server NEEDS: gnumake381,jtreg,scimark linux_i586-product-c1-scimark success( 42s) USED: hostname=sc11137358 platform=linux_x64_3.2 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=7890MB instance=P1 TIMING: clean=2s init=7s work=30s fini=1s VMFLAGS: -client NEEDS: gnumake381,jtreg,scimark linux_i586-product-c2-scimark success( 36s) USED: hostname=sc11137359 platform=linux_x64_3.2 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=7890MB instance=P1 TIMING: clean=1s init=7s work=27s fini=1s VMFLAGS: -server NEEDS: gnumake381,jtreg,scimark linux_x64-fastdebug-c2-scimark success( 42s) USED: hostname=sca00afg platform=linux_x64_2.6 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=16089MB instance=P1 TIMING: clean=2s init=6s work=32s fini=1s VMFLAGS: -server NEEDS: gnumake381,jtreg,scimark linux_x64-product-c2-scimark success( 39s) USED: hostname=sc11137360 platform=linux_x64_3.2 osname=linux osarch=x64 cpus=4 parallelcount=4 ram=7890MB instance=P1 TIMING: clean=2s init=10s work=26s fini=1s VMFLAGS: -server NEEDS: gnumake381,jtreg,scimark macosx_x64-fastdebug-c2-scimark success( 45s) USED: hostname=sc21jprtmac50 platform=macosx_x64_10.7 osname=macosx osarch=x64 cpus=16 parallelcount=16 instance=P1 TIMING: clean=1s init=10s work=33s fini=1s VMFLAGS: -server NEEDS: gnumake381,jtreg,scimark macosx_x64-product-c2-scimark success( 42s) USED: hostname=sc21jprtmac04 platform=macosx_x64_10.8 osname=macosx osarch=x64 cpus=4 parallelcount=4 instance=P1 TIMING: clean=1s init=8s work=31s fini=1s VMFLAGS: -server NEEDS: gnumake381,jtreg,scimark solaris_sparcv9-fastdebug-c2-scimark success( 50s) USED: hostname=sca00adr platform=solaris_sparc_5.10 osname=solaris osarch=sparc cpus=6 parallelcount=6 ram=31900MB instance=P1 TIMING: clean=3s init=10s work=36s fini=1s VMFLAGS: -d64 -server NEEDS: SS121,gnumake381,jtreg,scimark solaris_sparcv9-product-c2-scimark success(01m 22s) USED: hostname=sc11136044 platform=solaris_sparc_5.10 osname=solaris osarch=sparc cpus=128 parallelcount=32 ram=65312MB instance=P1 TIMING: clean=4s init=28s work=47s fini=1s VMFLAGS: -d64 -server NEEDS: SS121,gnumake381,jtreg,scimark solaris_x64-fastdebug-c2-scimark success( 45s) USED: hostname=sca00acl platform=solaris_i586_5.10 osname=solaris osarch=i586 cpus=6 parallelcount=6 ram=124000MB instance=T1 TIMING: clean=1s init=10s work=33s fini=1s VMFLAGS: -d64 -server NEEDS: SS121,gnumake381,jtreg,scimark solaris_x64-product-c2-scimark success( 39s) USED: hostname=sca00abe platform=solaris_i586_5.10 osname=solaris osarch=i586 cpus=6 parallelcount=6 ram=124000MB instance=T1 TIMING: clean=1s init=8s work=27s fini=1s VMFLAGS: -d64 -server NEEDS: SS121,gnumake381,jtreg,scimark windows_i586-fastdebug-c1-scimark success(01m 02s) USED: hostname=prt-win-x64-1 platform=windows_x64_5.2 osname=windows osarch=x64 cpus=4 parallelcount=4 ram=3967MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 mks=true cygwin=true installshield=true dxsdk=true pstools=true TIMING: clean=1s init=24s work=35s fini=1s VMFLAGS: -client NEEDS: cygwin,gnumake381,jtreg,scimark windows_i586-fastdebug-c2-scimark success(02m 48s) USED: hostname=sc11136490 platform=windows_x64_6.1 osname=windows osarch=x64 cpus=2 parallelcount=2 ram=7799MB instance=P1 compiler=VS2010 cygwin=true pstools=true TIMING: clean=1m50s init=21s work=35s fini=1s VMFLAGS: -server NEEDS: cygwin,gnumake381,jtreg,scimark windows_i586-product-c1-scimark success(01m 52s) USED: hostname=prt-win-x64-2 platform=windows_x64_5.2 osname=windows osarch=x64 cpus=4 parallelcount=4 ram=3967MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 mks=true cygwin=true installshield=true dxsdk=true pstools=true TIMING: clean=19s init=56s work=36s fini=1s VMFLAGS: -client NEEDS: cygwin,gnumake381,jtreg,scimark windows_i586-product-c2-scimark success(01m 50s) USED: hostname=prt-win-x64-1 platform=windows_x64_5.2 osname=windows osarch=x64 cpus=4 parallelcount=4 ram=3967MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 mks=true cygwin=true installshield=true dxsdk=true pstools=true TIMING: clean=16s init=58s work=33s fini=1s VMFLAGS: -server NEEDS: cygwin,gnumake381,jtreg,scimark windows_x64-fastdebug-c2-scimark success(02m 47s) USED: hostname=sca00aoh platform=windows_x64_6.1 osname=windows osarch=x64 cpus=6 parallelcount=6 ram=29999MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 cygwin=true pstools=true TIMING: clean=1m44s init=15s work=42s fini=1s VMFLAGS: -server NEEDS: cygwin,gnumake381,jtreg,scimark windows_x64-product-c2-scimark success(01m 16s) USED: hostname=sc11136119 platform=windows_x64_5.2 osname=windows osarch=x64 cpus=2 parallelcount=2 ram=8191MB instance=P1 compiler=VS2003,VS2005,VS2008,VS2010 mks=true cygwin=true installshield=true dxsdk=true pstools=true TIMING: clean=9s init=34s work=31s fini=1s VMFLAGS: -server NEEDS: cygwin,gnumake381,jtreg,scimark Platform Statistics linux_armsflt_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:47:25-PST N/A Tue-11:47:25-PST ended: Tue-12:17:00-PST N/A Tue-12:17:00-PST elapsed: 29m 35s N/A 29m 35s linux_armvfphflt_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:43:05-PST N/A Tue-11:43:05-PST ended: Tue-12:24:16-PST N/A Tue-12:24:16-PST elapsed: 41m 11s N/A 41m 11s linux_armvfpsflt_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:38:58-PST N/A Tue-11:38:58-PST ended: Tue-12:21:48-PST N/A Tue-12:21:48-PST elapsed: 42m 49s N/A 42m 49s linux_armvh_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:38:57-PST N/A Tue-11:38:57-PST ended: Tue-12:10:33-PST N/A Tue-12:10:33-PST elapsed: 31m 36s N/A 31m 36s linux_i586_2.6 (builds: 4) (tests: 4) (total) started: Tue-11:38:57-PST Tue-11:47:39-PST Tue-11:38:57-PST ended: Tue-12:08:41-PST Tue-12:14:39-PST Tue-12:14:39-PST elapsed: 29m 44s 27m 00s 35m 42s linux_ppc_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:38:57-PST N/A Tue-11:38:57-PST ended: Tue-12:00:48-PST N/A Tue-12:00:48-PST elapsed: 21m 51s N/A 21m 51s linux_ppcsflt_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:38:56-PST N/A Tue-11:38:56-PST ended: Tue-12:04:35-PST N/A Tue-12:04:35-PST elapsed: 25m 38s N/A 25m 38s linux_ppcv2_2.6 (builds: 2) (tests: 0) (total) started: Tue-11:38:20-PST N/A Tue-11:38:20-PST ended: Tue-12:04:20-PST N/A Tue-12:04:20-PST elapsed: 26m 00s N/A 26m 00s linux_x64_2.6 (builds: 3) (tests: 2) (total) started: Tue-11:40:18-PST Tue-11:44:27-PST Tue-11:40:18-PST ended: Tue-11:51:20-PST Tue-11:52:05-PST Tue-11:52:05-PST elapsed: 11m 02s 07m 38s 11m 47s macosx_x64_10.7 (builds: 2) (tests: 2) (total) started: Tue-11:38:57-PST Tue-11:46:00-PST Tue-11:38:57-PST ended: Tue-11:46:13-PST Tue-11:46:58-PST Tue-11:46:58-PST elapsed: 07m 16s 58s 08m 01s solaris_sparcv9_5.10 (builds: 2) (tests: 2) (total) started: Tue-11:35:59-PST Tue-11:51:16-PST Tue-11:35:59-PST ended: Tue-11:57:16-PST Tue-11:58:40-PST Tue-11:58:40-PST elapsed: 21m 16s 07m 24s 22m 41s solaris_x64_5.10 (builds: 3) (tests: 2) (total) started: Tue-11:38:57-PST Tue-11:51:58-PST Tue-11:38:57-PST ended: Tue-11:53:51-PST Tue-14:49:11-PST Tue-14:49:11-PST elapsed: 14m 54s 02H 57m 13s 03H 10m 14s windows_i586_6.1 (builds: 2) (tests: 4) (total) started: Tue-11:38:57-PST Tue-11:46:06-PST Tue-11:38:57-PST ended: Tue-11:50:49-PST Tue-11:53:44-PST Tue-11:53:44-PST elapsed: 11m 52s 07m 37s 14m 47s windows_x64_6.1 (builds: 2) (tests: 2) (total) started: Tue-11:38:57-PST Tue-11:47:05-PST Tue-11:38:57-PST ended: Tue-11:48:18-PST Tue-11:51:05-PST Tue-11:51:05-PST elapsed: 09m 21s 03m 59s 12m 07s For announcements on downtime and availability join the alias 'jprt_announce_ww_grp@oracle.com'. Status of the queue is available at: http://prt-web.us.oracle.com//archive//hotspotwest.html For suspected system problems, send mail to 'jprt_admin_ww_grp@oracle.com'. All jprt information is available at the JPRT wiki page http://wiki.se.oracle.com/display/JPGINFRA/JPRT
07-01-2014
Not all of the instances of this crash use DeoptimizeALot. How exactly do I rerun this same runThese test? Can I get the entire java command line including the testlist that it was using? I have my own runThese scripts that don't reproduce this problem. Is this run with ute or would I get the equivalent command with ute and -vmopts "<above options>" ?
07-01-2014
Yes, I can't find the fastdebug bundles! Maybe they are archived somewhere I don't know. As of now, I have nothing to work with.
03-01-2014
The JPRT build for the InfiniteStreamWithLimitOpTest failure: http://aurora-ds.us.oracle.com:9500/runs%2F323193.JAVASE.NIGHTLY.VM.Comp_Baseline-Tiered-Stable.2013-11-25-45/results/workDir/java/util/stream/test/org/openjdk/tests/java/util/stream/InfiniteStreamWithLimitOpTest Seems to be here: /net/prt-archiver.us.oracle.com/data/jprt/archive/2013/11/2013-11-22-220327.amurillo.hs25-b61-comp-sync/bundles Build the linux-x86 debug bundle, which created the core files, seems to have been deleted for file system space consideration :-(
03-01-2014
These InfiniteStreamWithLimitOpTest ones are even less reproduceable. Does anyone know how to find the binaries that produced the core files?
03-01-2014
This is really helpful. Thanks. There is a -client -Xcomp one that might be reproduceable and maybe the binaries exist also. I'm going to skip investigating the ppc64 one though.
02-01-2014
Here are the failure records in Aurora: http://aurora.ru.oracle.com/functional/faces/ChessBoard.xhtml?reportName=J2SEFailuresByReason&parameters=%5BreasonTypes%5DCrash%5BreasonDetails%5D%25must+be+alive%25%5BtestNameFilterRegExp%5D%25%5Bjdkproduct%5DJava%28TM%29%25%5Brelease%5D%25%5Bbuild%5D%25
02-01-2014
So the way this works is that the nmethod is unloaded if any of the oops in it aren't marked by GC nmethod::do_unloading. The class loader (instance of java.lang.ClassLoader) or mirror for anonymous classes are added to the oop section of the nmethod to keep metadata from being unloaded. In nmethod::do_unloading the nmethod is unloaded if any oops are not marked and then otherwise it does a verification that the metadata's loaders are alive. For this nmethod, there must be metadata whose holder (j.l.ClassLoader or mirror) isn't in the oop section. Or this checks the CompiledICHolders also. How that works is documented in wiki but I don't know where that page is. It seems to be gone.
02-01-2014
There isn't a wiki document about class unloading and their relation to nmethods. The metadata in nmethods are treated the same was as oops - they're weak references. The nmethod is unloaded if the oops/metadata contained in the nmethod is no longer alive. The logic is nmethod::do_unloading and can_unload is where we determine this. Never mind, this code always surprises me. I'll write something when I figure it out again. I can't find the binaries to match the core file and I can't reproduce this. The version that crashed is: Java HotSpot(TM) 64-Bit Server VM 25.0 b62 (25.0-b62-internal-201312021048.anoll.8028109-fastdebug) but it's not there.
02-01-2014
Moving to GC (or it should be runtime?) since the problem is not in compiled code but the state of a class loader which is referenced from compiled code. Do we have a document (wiki) which describes when and how class unloading happens and relationship with compiled methods (nmethod).
06-12-2013
Release team: Approved for deferral.
06-12-2013
DeoptimizeALot is non product flag, downgrading to P3
03-12-2013
vm_opts: -Xmixed -d64 -server -XX:MaxRAMFraction=8 -XX:+CreateMinidumpOnCrash -ea -esa -XX:-TieredCompilation -XX:CompileThreshold=100 -XX:+UnlockExperimentalVMOptions -XX:+IgnoreUnrecognizedVMOptions -XX:+DeoptimizeALot -XX:+AggressiveOpts version: jdk8 b118 / hs25.0 b62
03-12-2013