APAR status |
Closed as program error.
| Error description
This crash occurs when executing a user method that has been
jit compiled. In the specific problem that was observed, the
user had jdk ca122-20011213was.
.
The javacore displayed the following data
SIGSEGV received at 324706e0 in . Processing terminated.
.
Customer had the following jvm settings. It is not clear if
these contributed in anyway to the jvm crash, but these are
non-default.
-Xmx512M
-Xms256M
-Xcompactgc
-Xoss819200
-Xss819200
-Xnoclassgc
.
The java stack had this as the topmost call
Current Thread Details
----------------------
"Worker#12" (TID:0x8122fcd0, sys_thread_t:0x31574878,
state:R, native ID:0x1c1e) prio=5
at java.lang.reflect.Method.invoke(Native Method)
.
----- Native Stack -----
<allegedly at 0x324706e0 (not in text area)>
.
The jcore output shows the following details in the call stack:
SIGILL raised in libjitc.a
.
(cc)=4800001C at .
unwinding (old_sp=324706E0 old_ip=4800001C) old_mb=00000000
(Note that the method name not being identified may be due to
the fact that the core was truncated)
.
(pc)=35F5D5CF at
jp/co/dandi/trust/math/model/main/MG116Model.controlProc
(Ljp/co/dandi/trust/math/common/real/MathBean;)Z
unwinding (old_sp=3247070C old_ip=30280040) old_mb=35F5BC78
(pc)=35F5DE66 at
jp/co/dandi/trust/math/model/main/MG116Model.exec
(Ljp/co/dandi/trust/math/common/real/MathBean;)V
unwinding (old_sp=32470748 old_ip=30280040) old_mb=35F5BF78
(pc)=35ED8254 at
jp/co/dandi/trust/math/common/model/MathAPLController.exec
(Ljp/co/dandi/sysbase/real/SysbaseBean;)V
unwinding (old_sp=32470778 old_ip=302800E0) old_mb=35ED8068
(pc)=35E748F6 at
jp/co/dandi/trust/math/common/real/MathBean.exec ()V
unwinding (old_sp=324707B8 old_ip=30280020) old_mb=35E73FB8
(pc)=35EE4B5C at
jp/co/dandi/trust/math/real/MG116Bean.doFunctionModify ()V
unwinding (old_sp=324707E8 old_ip=30280020) old_mb=35EE44E0
(pc)=35EE4AF5 at
jp/co/dandi/trust/math/real/MG116Bean.doFunctionInput ()V
unwinding (old_sp=32470814 old_ip=30280020) old_mb=35EE4420
unwinding (old_sp=32470840 old_ip=30280180)
(ip)=D473B670 at libjvm.a .mmipExecuteJava + 0458
unwinding (old_sp=32470AC8 old_ip=D473B670)
(ip)=D47641C8 at libjvm.a .xeRunJvmMethod + 0124
unwinding (old_sp=32470B58 old_ip=D47641C8)
(ip)=D471C950 at libjvm.a .JVM_InvokeMethod + 0130
unwinding (old_sp=32470BC8 old_ip=D471C950)
(ip)=D47F1B6C at libjava.a
.Java_java_lang_reflect_Method_invoke + 0010
unwinding (old_sp=32470C08 old_ip=D47F1B6C)
.
The jit debugging that the customer did shows that the
NQOPTIMIZE step of the jit compiling that is causing the jit
crash. When the customer followed the jit debugging steps and
disabled this part of the jdk, the problem did not occur.
. The jcore output shows the following details in the call stack:SIGILL raised in libjitc.a.(cc)=4800001C at .unwinding (old_sp=324706E0 old_ip=4800001C) old_mb=00000000(Note that the method name not being identified may be due tothe fact that the core was truncated).(pc)=35F5D5CF atjp/co/dandi/trust/math/model/main/MG116Model.controlProc(Ljp/co/dandi/trust/math/common/real/MathBean;)Zunwinding (old_sp=3247070C old_ip=30280040) old_mb=35F5BC78(pc)=35F5DE66 atjp/co/dandi/trust/math/model/main/MG116Model.exec(Ljp/co/dandi/trust/math/common/real/MathBean;)Vunwinding (old_sp=32470748 old_ip=30280040) old_mb=35F5BF78(pc)=35ED8254 atjp/co/dandi/trust/math/common/model/MathAPLController.exec(Ljp/co/dandi/sysbase/real/SysbaseBean;)Vunwinding (old_sp=32470778 old_ip=302800E0) old_mb=35ED8068(pc)=35E748F6 atjp/co/dandi/trust/math/common/real/MathBean.exec ()Vunwinding (old_sp=324707B8 old_ip=30280020) old_mb=35E73FB8(pc)=35EE4B5C atjp/co/dandi/trust/math/real/MG116Bean.doFunctionModify ()Vunwinding (old_sp=324707E8 old_ip=30280020) old_mb=35EE44E0(pc)=35EE4AF5 atjp/co/dandi/trust/math/real/MG116Bean.doFunctionInput ()Vunwinding (old_sp=32470814 old_ip=30280020) old_mb=35EE4420unwinding (old_sp=32470840 old_ip=30280180)(ip)=D473B670 at libjvm.a .mmipExecuteJava + 0458unwinding (old_sp=32470AC8 old_ip=D473B670)(ip)=D47641C8 at libjvm.a .xeRunJvmMethod + 0124unwinding (old_sp=32470B58 old_ip=D47641C8)(ip)=D471C950 at libjvm.a .JVM_InvokeMethod + 0130unwinding (old_sp=32470BC8 old_ip=D471C950)(ip)=D47F1B6C at libjava.a.Java_java_lang_reflect_Method_invoke + 0010unwinding (old_sp=32470C08 old_ip=D47F1B6C).The jit debugging that the customer did shows that theNQOPTIMIZE step of the jit compiling that is causing the jitcrash. When the customer followed the jit debugging steps anddisabled this part of the jdk, the problem did not occur.. Local fix
This problem has been identified as sov defect 63367. It is
fixed in ca122-20030903 Problem summary
****************************************************************
* USERS AFFECTED: All WebSphere Application Server users of *
* the IBM Developer Kit, Java Technology *
* Edition prior to version ca122-20030903 . *
****************************************************************
* PROBLEM DESCRIPTION: The Just-In-Time ( JIT ) Compiler *
* crashes due to a SIGILL signal raised *
* in the libjitc.a file of the *
* application server Java Virtual Machine *
* ( JVM ). *
****************************************************************
* RECOMMENDATION: *
****************************************************************
The crash occurs when executing a user method that has been
JIT compiled. In the specific problem that was observed, the
user had the IBM Developer Kit version ca122-20011213was.
The javacore displayed the following data:
SIGSEGV received at 324706e0 in .
Processing terminated.
The user also had the following JVM settings. It is not clear
if these contributed in anyway to the JVM crash, but these are
not the default settings:The javacore displayed the following data:SIGSEGV received at 324706e0 in .Processing terminated.The user also had the following JVM settings. It is not clearif these contributed in anyway to the JVM crash, but these are
-Xmx512M
-Xms256M
-Xcompactgc
-Xoss819200
-Xss819200
-Xnoclassgc
The java stack had this as the topmost call
Current Thread Details
----------------------
"Worker#12" (TID:0x8122fcd0, sys_thread_t:0x31574878,
state:R, native ID:0x1c1e) prio=5
at java.lang.reflect.Method.invoke(Native Method)
----- Native Stack -----<allegedly at 0x324706e0 (not in text area)>
The jcore output shows the following details in the call stack:not the default settings:-Xmx512M-Xms256M-Xcompactgc-Xoss819200-Xss819200-XnoclassgcThe java stack had this as the topmost callCurrent Thread Details----------------------"Worker#12" (TID:0x8122fcd0, sys_thread_t:0x31574878,state:R, native ID:0x1c1e) prio=5at java.lang.reflect.Method.invoke(Native Method)----- Native Stack -----
SIGILL raised in libjitc.a
(cc)=4800001C at .
unwinding (old_sp=324706E0 old_ip=4800001C) old_mb=00000000
(Note that the method name not being identified may be due to
the fact that the core was truncated)
(pc)=35F5D5CF at
jp/co/dandi/trust/math/model/main/MG116Model.controlProc
(Ljp/co/dandi/trust/math/common/real/MathBean;)Z
unwinding (old_sp=3247070C old_ip=30280040) old_mb=35F5BC78
(pc)=35F5DE66 at
jp/co/dandi/trust/math/model/main/MG116Model.exec
(Ljp/co/dandi/trust/math/common/real/MathBean;)V
unwinding (old_sp=32470748 old_ip=30280040) old_mb=35F5BF78
(pc)=35ED8254 at
jp/co/dandi/trust/math/common/model/MathAPLController.exec
(Ljp/co/dandi/sysbase/real/SysbaseBean;)V
unwinding (old_sp=32470778 old_ip=302800E0) old_mb=35ED8068
(pc)=35E748F6 at
jp/co/dandi/trust/math/common/real/MathBean.exec ()V
unwinding (old_sp=324707B8 old_ip=30280020) old_mb=35E73FB8
(pc)=35EE4B5C at
jp/co/dandi/trust/math/real/MG116Bean.doFunctionModify ()V
unwinding (old_sp=324707E8 old_ip=30280020) old_mb=35EE44E0
(pc)=35EE4AF5 at
jp/co/dandi/trust/math/real/MG116Bean.doFunctionInput ()V
unwinding (old_sp=32470814 old_ip=30280020) old_mb=35EE4420
unwinding (old_sp=32470840 old_ip=30280180)
(ip)=D473B670 at libjvm.a .mmipExecuteJava + 0458
unwinding (old_sp=32470AC8 old_ip=D473B670)
(ip)=D47641C8 at libjvm.a .xeRunJvmMethod + 0124
unwinding (old_sp=32470B58 old_ip=D47641C8)
(ip)=D471C950 at libjvm.a .JVM_InvokeMethod + 0130
unwinding (old_sp=32470BC8 old_ip=D471C950)
(ip)=D47F1B6C at libjava.a
.Java_java_lang_reflect_Method_invoke + 0010
unwinding (old_sp=32470C08 old_ip=D47F1B6C)
The JIT debugging shows that the NQOPTIMIZE step of the JIT
compiling is causing the JIT to crash. The user followed the
JIT debugging steps and disabled this part of the IBM Developer
Kit, the problem did not occur. The jcore output shows the following details in the call stack:SIGILL raised in libjitc.a(cc)=4800001C at .unwinding (old_sp=324706E0 old_ip=4800001C) old_mb=00000000(Note that the method name not being identified may be due tothe fact that the core was truncated)(pc)=35F5D5CF atjp/co/dandi/trust/math/model/main/MG116Model.controlProc(Ljp/co/dandi/trust/math/common/real/MathBean;)Zunwinding (old_sp=3247070C old_ip=30280040) old_mb=35F5BC78(pc)=35F5DE66 atjp/co/dandi/trust/math/model/main/MG116Model.exec(Ljp/co/dandi/trust/math/common/real/MathBean;)Vunwinding (old_sp=32470748 old_ip=30280040) old_mb=35F5BF78(pc)=35ED8254 atjp/co/dandi/trust/math/common/model/MathAPLController.exec(Ljp/co/dandi/sysbase/real/SysbaseBean;)Vunwinding (old_sp=32470778 old_ip=302800E0) old_mb=35ED8068(pc)=35E748F6 atjp/co/dandi/trust/math/common/real/MathBean.exec ()Vunwinding (old_sp=324707B8 old_ip=30280020) old_mb=35E73FB8(pc)=35EE4B5C atjp/co/dandi/trust/math/real/MG116Bean.doFunctionModify ()Vunwinding (old_sp=324707E8 old_ip=30280020) old_mb=35EE44E0(pc)=35EE4AF5 atjp/co/dandi/trust/math/real/MG116Bean.doFunctionInput ()Vunwinding (old_sp=32470814 old_ip=30280020) old_mb=35EE4420unwinding (old_sp=32470840 old_ip=30280180)(ip)=D473B670 at libjvm.a .mmipExecuteJava + 0458unwinding (old_sp=32470AC8 old_ip=D473B670)(ip)=D47641C8 at libjvm.a .xeRunJvmMethod + 0124unwinding (old_sp=32470B58 old_ip=D47641C8)(ip)=D471C950 at libjvm.a .JVM_InvokeMethod + 0130unwinding (old_sp=32470BC8 old_ip=D471C950)(ip)=D47F1B6C at libjava.a.Java_java_lang_reflect_Method_invoke + 0010unwinding (old_sp=32470C08 old_ip=D47F1B6C)The JIT debugging shows that the NQOPTIMIZE step of the JITcompiling is causing the JIT to crash. The user followed theJIT debugging steps and disabled this part of the IBM DeveloperKit, the problem did not occur. Problem conclusion
This problem has been identified as IBM Developer Kit sov defect
63367. It is fixed in version ca122-20030903 of the IBM
Developer Kit. Temporary fixComments
APAR information | APAR number | PQ78069 | Reported component name | WAS STANDARD AI | Reported component ID | 5648C8300 | Reported release | 350 | Status | CLOSED PER | PE | NoPE | HIPER | NoHIPER | Submitted date | 2003-09-03 | Closed date | 2003-09-12 | Last modified date | 2003-09-12 |
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:APAR is sysrouted FROM one or more of the following:
Modules/Macros APAR is sysrouted TO one or more of the following:Modules/Macros
Applicable component levels | R350 PSY | UP |
|