is this a bug ?

Yangfei (Felix) felix.yang at huawei.com
Wed Aug 14 06:56:20 UTC 2019


Hi,

We witnessed random fail for this test on our 64/128 core aarch64 server platform: org.openjdk.jcstress.tests.atomicity.buffers.DirectByteBufferViewsAtomicityTests.LongViewTest

After some investigation, we found one bug was introduced in this commit: http://hg.openjdk.java.net/code-tools/jcstress/rev/8c4495a44ffa

Patch is trivial:
diff -r 7dce6af27fd5 tests-custom/src/main/java/org/openjdk/jcstress/tests/atomicity/buffers/DirectByteBufferViewsAtomicityTests.java
--- a/tests-custom/src/main/java/org/openjdk/jcstress/tests/atomicity/buffers/DirectByteBufferViewsAtomicityTests.java  Sun Aug 11 10:37:40 2019 +0200
+++ b/tests-custom/src/main/java/org/openjdk/jcstress/tests/atomicity/buffers/DirectByteBufferViewsAtomicityTests.java  Thu Aug 15 14:42:49 2019 +0800
@@ -52,7 +52,7 @@
         private final ShortBuffer sb;

         public MyState() {
-            b = ByteBuffer.allocate(16);
+            b = ByteBuffer.allocateDirect(16);
             b.order(ByteOrder.nativeOrder());
             ib = b.asIntBuffer();
             cb = b.asCharBuffer();

Please confirm.

Thanks for your help,
Felix


More information about the code-tools-dev mailing list