changeset in /hg/icedtea6: 2009-01-26 Gary Benson <gbenson at red...

[Gary Benson]

changeset 2b1c03c1e9fa in /hg/icedtea6
details: http://icedtea.classpath.org/hg/icedtea6?cmd=changeset;node=2b1c03c1e9fa
description:
	2009-01-26  Gary Benson  <gbenson at redhat.com>

		* ports/hotspot/src/os_cpu/linux_zero/vm/thread_linux_zero.hpp
		(JavaThread::has_special_condition_for_native_trans): New method.
		* ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp:
		(CppInterpreter::native_entry): Correctly handle safepoints
		during post-call thread state transition.

diffstat:

3 files changed, 21 insertions(+), 8 deletions(-)
ChangeLog                                                    |    8 ++++++
ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp        |   12 +++-------
ports/hotspot/src/os_cpu/linux_zero/vm/thread_linux_zero.hpp |    9 +++++++

diffs (56 lines):

diff -r fed3590e18e7 -r 2b1c03c1e9fa ChangeLog
--- a/ChangeLog	Thu Jan 22 15:06:43 2009 -0500
+++ b/ChangeLog	Mon Jan 26 08:36:15 2009 -0500
@@ -1,3 +1,11 @@ 2009-01-22  Gary Benson  <gbenson at redhat
+2009-01-26  Gary Benson  <gbenson at redhat.com>
+
+	* ports/hotspot/src/os_cpu/linux_zero/vm/thread_linux_zero.hpp
+	(JavaThread::has_special_condition_for_native_trans): New method.
+	* ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp:
+	(CppInterpreter::native_entry): Correctly handle safepoints
+	during post-call thread state transition.
+
 2009-01-22  Gary Benson  <gbenson at redhat.com>
 
 	* ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp:
diff -r fed3590e18e7 -r 2b1c03c1e9fa ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp
--- a/ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp	Thu Jan 22 15:06:43 2009 -0500
+++ b/ports/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp	Mon Jan 26 08:36:15 2009 -0500
@@ -315,14 +315,10 @@ void CppInterpreter::native_entry(method
     }
   }
 
-  // We need to call check_special_condition_for_native_trans() if
-  // any of the thread's suspend flags are set.  We can't use the
-  // various accessors because thread->_suspend_flags is volatile
-  // and may change while we're reading it.
-  volatile uint32_t *suspend_flags;
-  suspend_flags = (uint32_t *)
-    ((address) thread + in_bytes(JavaThread::suspend_flags_offset()));
-  if (*suspend_flags != 0) {
+  // Handle safepoint operations, pending suspend requests,
+  // and pending asynchronous exceptions.
+  if (SafepointSynchronize::do_call_back() ||
+      thread->has_special_condition_for_native_trans()) {
     JavaThread::check_special_condition_for_native_trans(thread);
     CHECK_UNHANDLED_OOPS_ONLY(thread->clear_unhandled_oops());
   }
diff -r fed3590e18e7 -r 2b1c03c1e9fa ports/hotspot/src/os_cpu/linux_zero/vm/thread_linux_zero.hpp
--- a/ports/hotspot/src/os_cpu/linux_zero/vm/thread_linux_zero.hpp	Thu Jan 22 15:06:43 2009 -0500
+++ b/ports/hotspot/src/os_cpu/linux_zero/vm/thread_linux_zero.hpp	Mon Jan 26 08:36:15 2009 -0500
@@ -95,6 +95,15 @@
   }
 
  public:
+  // Check for pending suspend requests and pending asynchronous
+  // exceptions.  There are separate accessors for these, but
+  // _suspend_flags is volatile so using them would be unsafe.
+  bool has_special_condition_for_native_trans()
+  {
+    return _suspend_flags != 0;
+  }
+
+ public:
   bool pd_get_top_frame_for_signal_handler(frame* fr_addr,
                                            void* ucontext,
                                            bool isInJava)