JDK-6232217 : Regression: test for failing for 1.4.2_08 b02 but passes for 1.4.2_08 b01 on Cinnebar
  • Type: Bug
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 1.4.2_08
  • Priority: P2
  • Status: Closed
  • Resolution: Duplicate
  • OS: linux
  • CPU: x86
  • Submitted: 2005-02-23
  • Updated: 2010-05-04
  • Resolved: 2005-03-01
Related Reports
Duplicate :  
JDK-6231558 - JVM outputs exceptions or crash with a program using Big interger in 1.4.2_08b2
Relates :  
JDK-6215242 - ShouldNotReachHere() failure at hotspot/src/cpu/i486/vm/c1_FrameMap_i486.cpp, 92
Relates :  
JDK-4917709 - SEGV in MapLoops test
Description
When running the regression tests for 1.4.2_08 b2 the test for 4917709 fails on
Cinnebar. The same result is seen when the test is run manually.
This test PASSES for 1.4.2_08 b01.

The test case can be found at:
/net/latte.ireland/deployment1/regtests/142/test/CTE_REGTEST/Generic/4917709

To reproduce copy Tester.java from the above location, compile and run it.

The jtr file is attached.
###@###.### 2005-2-23 16:15:20 GMT
###@###.### 2005-2-24 13:33:44 GMT
Comments
EVALUATION This bug has the same root cause as 4917709. The fix for 6215242 inadvertently reintroduced 4917709. A better fix is to reintroduce the bailout that was present in the 1.4.0 sources; removing this bailout in 1.4.1 caused 4917709. This doesn't affect the 1.5 source tree due to the introduction of safepoint polling and the removal of safepoints from the backward branch instruction. This bug will probably be closed as a duplicate of 6231558 once that bug is marked fixed. ###@###.### 2005-2-25 02:46:12 GMT Closing per request of product team. 6231558 will be fixed in 1.4.2_08. ###@###.### 2005-03-01 23:03:06 GMT
25-02-2005