Crashes in Reference Chain Walker in jdwp and JVMTI tests

[dmitripivkine]

Discovered in internal testing:

Type=Segmentation error vmState=0x00000000
J9Generic_Signal_Number=00000018 Signal_Number=0000000b Error_Value=00000000 Signal_Code=00000001
Handler1=00007F955631DDB0 Handler2=00007F9555BC3CD0 InaccessibleAddress=000000003C08A430
RDI=0000000000000000 RSI=0000000000000000 RAX=000000003C08A400 RBX=00007F953C08A770
RCX=0000000000000011 RDX=0000000000000000 R8=FFFFFFFFFFFFFFFF R9=0000000000000001
R10=0000000000000003 R11=00000000FFE13A40 R12=00007F953C08A180 R13=00007F955000FDB0
R14=00007F953C08A2B0 R15=00000000FFE13A40
RIP=00007F954F86F369 GS=0000 FS=0000 RSP=00007F953C08A150
EFlags=0000000000010206 CS=0033 RBP=0000000000053E00 ERR=0000000000000004
TRAPNO=000000000000000E OLDMASK=0000000000000000 CR2=000000003C08A430
xmm0 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm1 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm2 408f400000000000 (f: 0.000000, d: 1.000000e+03)
xmm3 0000000000000789 (f: 1929.000000, d: 9.530526e-321)
xmm4 0000000042700000 (f: 1114636288.000000, d: 5.507035e-315)
xmm5 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm6 00000000be92618f (f: 3197264384.000000, d: 1.579658e-314)
xmm7 705f7366632e7570 (f: 1663989120.000000, d: 1.953116e+233)
xmm8 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm9 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm10 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm11 0000000000000000 (f: 0.000000, d: 0.000000e+00)
xmm12 ffffffff00000000 (f: 0.000000, d: -nan)
xmm13 110325c000000001 (f: 1.000000, d: 1.010333e-226)
xmm14 0000001c000012d8 (f: 4824.000000, d: 5.941588e-313)
xmm15 0701009a00079800 (f: 497664.000000, d: 6.138509e-275)
Module=/bluebird/builds/bld_60074/sdk/xa6480/jre/lib/amd64/compressedrefs/libj9jvmti29.so
Module_base_address=00007F954F85B000
Target=2_90_20231101_60074 (Linux 4.12.14-122.179-default)
CPU=amd64 (4 logical CPUs) (0x1e3420000 RAM)
----------- Stack Backtrace -----------
followReferencesCallback+0x1b9 (0x00007F954F86F369 [libj9jvmti29.so+0x14369])
_ZN23MM_ReferenceChainWalker6doSlotEPP8J9ObjectllS1_+0x9d (0x00007F955413E49D [libj9gc29.so+0x3249d])
walkPushedJNIRefs+0xc8 (0x00007F955634C568 [libj9vm29.so+0xfe568])
walkStackFrames+0x103c (0x00007F955634DEBC [libj9vm29.so+0xffebc])
_ZN28GC_VMThreadStackSlotIterator9scanSlotsEP10J9VMThreadS1_PvPFvP8J9JavaVMPP8J9ObjectS2_P16J9StackWalkStatePKvEbb+0x3c (0x00007F955414FADC [libj9gc29.so+0x43adc])
_ZN14MM_RootScanner13scanOneThreadEP18MM_EnvironmentBaseP10J9VMThreadPv+0xfd (0x00007F955414057D [libj9gc29.so+0x3457d])
_ZN14MM_RootScanner11scanThreadsEP18MM_EnvironmentBase+0xbf (0x00007F955413F28F [libj9gc29.so+0x3328f])
_ZN14MM_RootScanner12scanAllSlotsEP18MM_EnvironmentBase+0x38 (0x00007F9554142AD8 [libj9gc29.so+0x36ad8])
j9gc_ext_reachable_objects_do+0x1da (0x00007F955413E6FA [libj9gc29.so+0x326fa])
jvmtiFollowReferences+0x1e6 (0x00007F954F870816 [libj9jvmti29.so+0x15816])
objectReferrers+0x198 (0x00007F953D19C0B8 [libjdwp.so+0x2b0b8])
referringObjects+0xb4 (0x00007F953D176194 [libjdwp.so+0x5194])
debugLoop_run+0x1ba (0x00007F953D1826AA [libjdwp.so+0x116aa])
acceptThread+0x6f (0x00007F953D1947FF [libjdwp.so+0x237ff])
wrappedAgentThreadStart+0xaf (0x00007F954F88336F [libj9jvmti29.so+0x2836f])
omrsig_protect+0x1e3 (0x00007F9555BC4A33 [libj9prt29.so+0x27a33])
agentThreadStart+0x68 (0x00007F954F8831C8 [libj9jvmti29.so+0x281c8])
thread_wrapper+0x186 (0x00007F9556041466 [libj9thr29.so+0xe466])
start_thread+0xca (0x00007F955814871A [libpthread.so.0+0x971a])
clone+0x6d (0x00007F9557A6417D [libc.so.6+0xf017d])

Can it be regression from #18379 or #18378?

[dmitripivkine]

@LinHu2016 @babsingh @TobiAjila

[dmitripivkine]

@pshipton FYI

[JasonFengJ9]

This is tracked at RTC 150163

[pshipton]

@dmitripivkine since this is tracked internally, probably we don't need this open?

[pshipton]

Pls add the failure details internally, I'm closing this.

[dmitripivkine]

@dmitripivkine since this is tracked internally, probably we don't need this open?

My slight preference would be to keep this item, but we can use internal one if you like.