[15] RFR(XXS): 8247530: JfrCheckpointManager failed "assert(!SafepointSynchronize::is_at_safepoint()) failed: invariant"

Markus Gronlund markus.gronlund at oracle.com
Mon Jun 15 13:25:19 UTC 2020


Greetings,

Please review this change to remove two stale assertions (post JDK-8245113):

Summary:
The JFR Recorder Thread invoking clear() does not imply !SafepointSynchronize::is_at_safepoint() no more; there is no relation to safepointing to be asserted, because the thread has been excluded from the safepoint protocol at this location.

Bug: https://bugs.openjdk.java.net/browse/JDK-8247530
Change:

diff --git a/src/hotspot/share/jfr/recorder/checkpoint/jfrCheckpointManager.cpp b/src/hotspot/share/jfr/recorder/checkpoint/jfrCheckpointManager.cpp
--- a/src/hotspot/share/jfr/recorder/checkpoint/jfrCheckpointManager.cpp
+++ b/src/hotspot/share/jfr/recorder/checkpoint/jfrCheckpointManager.cpp
@@ -385,7 +385,6 @@
 }
 
 void JfrCheckpointManager::clear_type_set() {
-  assert(!SafepointSynchronize::is_at_safepoint(), "invariant");
   assert(!JfrRecorder::is_recording(), "invariant");
   Thread* t = Thread::current();
   DEBUG_ONLY(JfrJavaSupport::check_java_thread_in_native(t));
@@ -398,7 +397,6 @@
 }
 
 void JfrCheckpointManager::write_type_set() {
-  assert(!SafepointSynchronize::is_at_safepoint(), "invariant");
   {
     Thread* const thread = Thread::current();
     DEBUG_ONLY(JfrJavaSupport::check_java_thread_in_native(thread));


Thanks
Markus


More information about the hotspot-jfr-dev mailing list