I ran a round of Thread-SMR stress testing on the 2018-02-02
jdk/jdk10 snapshot on my Solaris X64 server.

The gtest/GTestWrapper.java test failed opto.mathexact_test_vm
in both the 'release' and 'fastdebug' configs.

Here's the snippet from the 'release' log:

[----------] 1 test from opto
[ RUN ] opto.mathexact_test_vm
/work/shared/bug_hunt/8167108/SMR_prototype_10/open/test/hotspot/gtest/opto/test_mathexact.cpp:34: Failure
Value of: OverflowMulLNode::is_overflow(-1, max_jlong)
  Actual: true
Expected: false
[ FAILED ] opto.mathexact_test_vm (0 ms)
[----------] 1 test from opto (0 ms total)

Here's the snippet from the 'fastdebug' log:

[----------] 1 test from opto
[ RUN ] opto.mathexact_test_vm
/work/shared/bug_hunt/8167108/SMR_prototype_10/open/test/hotspot/gtest/opto/test_mathexact.cpp:34: Failure
Value of: OverflowMulLNode::is_overflow(-1, max_jlong)
  Actual: true
Expected: false
[ FAILED ] opto.mathexact_test_vm (0 ms)
[----------] 1 test from opto (0 ms total)