IJ35231: INCORRECT RESULT TRANSFORMING BYTE TO CHAR IN LOOPS FOR LITTLE ENDIAN

APAR status

• Closed as program error.

Error description

• Error Message: N/A
  .
  Stack Trace: N/A
  .
  Incorrect or unexpected values are returned when transforming
  byte to char in a loop. The issue happens more when compiled at
  opt level hot or higher.
  The problem happens only on little endian machines, and cannot
  happen on big endian machines like AIX and zOS.
  

Local fix

• The issue can be avoided with -Xjit:disableIdiomRecognition.
  This option has some performance impact, but it may be
  restricted to the problematic method ex:
  -Xjit:{<effected_method_signature>}(disableIdiomRecognition)
  

Problem summary

• The problem happens in the Idiom recognition optimization when
  it transforms the loop into an arraycopy. The optimization does
  a wrong calculation on little endian architectures for the
  starting address for arraycopy.
  

Problem conclusion

• The compiler code has been corrected to select the starting
  address correctly based on the endianess of the target platform.
  .
  This APAR will be fixed in the following Java Releases:
     8    SR7 FP5   (8.0.7.5)
  .
  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