IJ23414: UNEXPECTED BEHAVIOUR: HASH MAP RETURNING INCORRECT HASH CODE VALUE AFTER CERTAIN ITERATION ON LINUX ON Z

APAR status

• Closed as program error.

Error description

• Error Message: N/A
  .
  Stack Trace:
  .
  

Local fix

• N/A
  

Problem summary

• There was a bug in a compiler code which generates an optimized
  compare and branch instruction on IBM Z codegen when we have a
  code that compares class field of two objects. Class field in
  object on 64-Bit platform with compressed reference is treated
  specially as it only takes up 4 bytes in memory. If one of the
  class field is already used in same method earlier, there are
  good chances that instead of loading it again from an object, we
  would temporary store it on Java Stack which takes up 8 bytes.
  Compiler was generating wrong code that only compares 4 bytes.
  This is incorrect on 64-Bit platform as the address load from
  temp slot on the java stack takes up 8 bytes in the memory.
  

Problem conclusion

• Compiler code has been updated to identify such cases and
  generate code correct data type for the operation.
  .
  This APAR will be fixed in the following Java Releases:
     8    SR6 FP7   (8.0.6.7)
     7    SR10 FP65 (7.0.10.65)
     7 R1 SR4 FP65  (7.1.4.65)
  .
  Contact your IBM Product's Service Team for these Service
  Refreshes and Fix Packs.
  For those running stand-alone, information about the available
  Service Refreshes and Fix Packs can be found at:
             https://www.ibm.com/developerworks/java/jdk/
  

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

  JIT

• Fixed component ID

  620700124

Applicable component levels