Open liqunl opened 4 years ago
Grinder from @rpshukla in #10623
10x grinder for MauveSingleInvocationLoadTest_special_22 passed: https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3951/
running more grinders: https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3952/ https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3953/
This grinder https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3953/ was supposed to be for the original failure in #10623 (although it looks like I didn't get the settings correct and no tests were actually run)
The others are for MauveSingleInvocationLoadTest_special_22
Haven't seen a failure yet. Here are a couple more grinders:
https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3954/ 20/20 passed https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3955/ 20/20 passed https://ci.adoptopenjdk.net/view/Test_grinder/job/Grinder/3956/ 20/20 passed
Could be related to #10365, will update with my finding later.
It seems to be different than #10365, there are different problems show up in the core dump. It is a crash in populateVPicSlotCall, when we're trying to get j9method from vtable of a class. The object reference passed to the helper is bad, thus the j9class we got from it is also bad. Looking at its caller, there are the following problems
I'll post more detailed investigation in another comment later.
I haven't figured out why so many things are wrong at the same time. Given the options -Xjit:count=0 -Xgcpolicy:gencon -Xaggressive -Xconcurrentlevel0
and that we can't reproduce it in 80 runs, I would recommend we defer it to next release. @andrewcraik
I haven't made any progress on this issue. Given the time left and other issues I'm working on, I suggest we defer this to next release
@liqunl I'm not sure why this is in the milestone plan at all. It's not a failure I see in the nightly builds so it can't be happening often. I suspect it was added only because https://github.com/eclipse/openj9/issues/10623 was in the milestone plan. Unless you have some reason to keep it in, I'll remove it rather than moving it forward.
OK, I think we can remove the milestone target.
https://ci.eclipse.org/openj9/job/Test_openjdk8_j9_special.system_x86-64_mac_mixed_Nightly_testList_0/5 MauveSingleInvocLoad_special_5m_22 variation: Mode688 JVM_OPTIONS: -Xcompressedrefs -Xjit:count=0 -Xgcpolicy:gencon -Xaggressive -Xconcurrentlevel0 https://140-211-168-230-openstack.osuosl.org/artifactory/ci-eclipse-openj9/Test/Test_openjdk8_j9_special.system_x86-64_mac_mixed_Nightly_testList_0/5/system_test_output.tar.gz
LT stderr Unhandled exception
LT stderr Type=Segmentation error vmState=0x00000000
LT stderr J9Generic_Signal_Number=00000018 Signal_Number=0000000b Error_Value=00000000 Signal_Code=00000001
LT stderr Handler1=0000000004036240 Handler2=0000000003798860 InaccessibleAddress=0000000000000019
LT stderr RDI=000000001D1640B3 RSI=000000001D1640B8 RAX=0000000000000001 RBX=00000000FFFEF708
LT stderr RCX=00000000FFB59C00 RDX=000000001D1640B8 R8=00000000FFB59C00 R9=00000000FFFEEC20
LT stderr R10=0000000080AEA1E0 R11=00007F80E0003890 R12=0000000000000000 R13=00000000FFFEF728
LT stderr R14=00000000FFFEF6B8 R15=00000000FFFEF6C8
LT stderr RIP=000000000476F668 GS=0000 FS=0000 RSP=000000000FD28F80
LT stderr RFlags=0000000000010203 CS=002B RBP=000000000FD21900 ERR=0000001900000004
LT stderr TRAPNO=000000040000000E CPU=0019000000040000 FAULTVADDR=0000000000000019
LT stderr XMM0 0000000000000002 (f: 2.000000, d: 9.881313e-324)
LT stderr XMM1 000000000fd28ff8 (f: 265457664.000000, d: 1.311535e-315)
LT stderr XMM2 00000000df000000 (f: 3741319168.000000, d: 1.848457e-314)
LT stderr XMM3 3e12930f252766de (f: 623339200.000000, d: 1.081175e-09)
LT stderr XMM4 3eb26186b931f159 (f: 3107057920.000000, d: 1.095591e-06)
LT stderr XMM5 00000000003c0000 (f: 3932160.000000, d: 1.942745e-317)
LT stderr XMM6 3fd947941c2116fb (f: 471930624.000000, d: 3.949938e-01)
LT stderr XMM7 40262e42fefa39ef (f: 4277811712.000000, d: 1.109035e+01)
LT stderr XMM8 0000000041800000 (f: 1098907648.000000, d: 5.429325e-315)
LT stderr XMM9 0000000000000000 (f: 0.000000, d: 0.000000e+00)
LT stderr XMM10 0000000000000006 (f: 6.000000, d: 2.964394e-323)
LT stderr XMM11 0000000000000008 (f: 8.000000, d: 3.952525e-323)
LT stderr XMM12 bf9f518686afffa1 (f: 2259681280.000000, d: -3.058443e-02)
LT stderr XMM13 3f9431defbd6fb77 (f: 4225170176.000000, d: 1.972149e-02)
LT stderr XMM14 bf705fca77713fdd (f: 2003910656.000000, d: -3.997603e-03)
LT stderr XMM15 3f14d74035fec9a7 (f: 905890240.000000, d: 7.950143e-05)
LT stderr Module=/Users/jenkins/workspace/Test_openjdk8_j9_special.system_x86-64_mac_mixed_Nightly_testList_0/openjdkbinary/j2sdk-image/jre/lib/default/libj9jit29.dylib
LT stderr Module_base_address=00000000044AB000 Symbol=..@109.done
LT stderr Symbol_address=000000000476F650
LT stderr Target=2_90_20210207_6 (Mac OS X 10.13.6)
LT stderr CPU=amd64 (4 logical CPUs) (0x200000000 RAM)
LT stderr ----------- Stack Backtrace -----------
LT stderr ---------------------------------------
@liqunl do recent updates to the jitdump help with his?
https://ci.eclipse.org/openj9/job/Test_openjdk8_j9_special.system_x86-64_windows_Nightly_mauveLoadTest/134 MauveSingleInvocationLoadTest_special_22 Options are:
-Xcompressedrefs -Xjit:count=0 -Xgcpolicy:gencon -Xaggressive -Xconcurrentlevel0