[Bug 500] Shark on ARM/x86_64 Math.log incosistent results

[bugzilla-daemon at icedtea.classpath.org]

http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=500





------- Comment #3 from xerxes at zafena.se  2010-05-25 10:57 -------
By using the following patch makes ARM pass the testcase by preventing Shark to
use the llvm intrinsics for all aproximative math function and instead only
compile the jni wrapper to the same math intrinsic used by the interpreter.

Index: openjdk/hotspot/src/share/vm/shark/sharkIntrinsics.cpp
===================================================================
--- openjdk.orig/hotspot/src/share/vm/shark/sharkIntrinsics.cpp 2010-05-25
12:02:16.000000000 +0200
+++ openjdk/hotspot/src/share/vm/shark/sharkIntrinsics.cpp      2010-05-25
12:02:29.000000000 +0200
@@ -46,7 +46,7 @@
   case vmIntrinsics::_dlog10:
   case vmIntrinsics::_dpow:
   case vmIntrinsics::_dexp:
-    return true;
+    return false;

     // java.lang.Object
   case vmIntrinsics::_getClass:


-- 
Configure bugmail: http://icedtea.classpath.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.