IV38666: GPF/ABORT IN MM_MEMORYSUBSPACESEMISPACE::CHECKSUBSPACEMEMORYPOS TCOLLECTRESIZE

APAR status

• Closed as program error.

Error description

• Error Message: A SIGSEGV/GPF is received during a GC
  .
  Stack Trace:
  MM_MemorySubSpaceSemiSpace::checkSubSpaceMemoryPostCollectRe
  MM_MemorySubSpaceSemiSpace::checkResize(MM_EnvironmentModron
  MM_Scavenger::masterThreadGarbageCollect(MM_EnvironmentModro
  MM_Scavenger::internalGarbageCollect(MM_EnvironmentModron*,M
  MM_ParallelScavenger::internalGarbageCollect(MM_EnvironmentM
  MM_Collector::garbageCollect(MM_EnvironmentModron*,MM_Memory
  MM_MemorySubSpaceSemiSpace::allocationRequestFailed(MM_Envir
  MM_MemorySubSpaceGeneric::allocateTLH(MM_EnvironmentModron*,
  MM_TLHAllocationInterface::refresh(MM_EnvironmentModron*,MM_
  MM_TLHAllocationInterface::allocateObject(MM_EnvironmentModr
  MM_TLHAllocationInterface::allocateArrayletSpine(MM_Environm
  MM_ArrayletAllocationModel::allocateAndInitializeLayout(MM_M
  MM_ArrayletAllocationModel::allocateIndexableObject(MM_Envir
  .
  Crash has only been observed on large systems with highly
  frequent scavenges and a small heap
  

Local fix

• Increase the size of the Heap to reduce the frequency of
  svacenges
  

Problem summary

• The problem is caused when the time between scavenges and the
  scan time are below 1ms, due to the accuracy of the system clock
  these numbers are rounded to 0 causing an IEEE
  exponent-underflow
  

Problem conclusion

• This defect will be fixed in:
  7.0.0 SR5
  6.0.1 SR6
  .
  The JVM has been updated to not resize the heap post collect if
  the time interval from the system is 0.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  J9 COMMON CODE

• Fixed component ID

  620700127

Applicable component levels

• R260 PSY

     UP