-
Bug
-
Resolution: Fixed
-
P3
-
11, 12, 13
-
b11
Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
---|---|---|---|---|---|---|
JDK-8220365 | 12.0.2 | Thomas Stuefe | P3 | Resolved | Fixed | b01 |
JDK-8226584 | 11.0.5-oracle | Thomas Stuefe | P3 | Resolved | Fixed | b02 |
JDK-8220654 | 11.0.4 | Thomas Stuefe | P3 | Resolved | Fixed | b01 |
JDK-8220433 | 11.0.3 | Thomas Stuefe | P3 | Resolved | Fixed | b03 |
The new test "test_print_hex_dump", introduced with JDK-8217994 has the issue, that it could cause access to an unaligned address. We've observed crashes on Solaris. Furthermore, the test should run as TEST_VM, because it uses VM features such as ResourceMark. This won't be a problem if the complete test suite is run but for standalone usage, it'll cause problems.
---
After discussing this, we thing that
- test should use TEST_VM
- the alignment problems are real issues and should be fixed in os::print_hex_dump();
---
Update: for the record, we had the following issues:
- Test did not use TEST_VM so signal handlers could be missing depending on test order
- os::print_hex_dump() had alignment problems.
- On Windows, the gtest will not work since SEH frame is missing. This is a general problem for gtests and I opened a separate issue for this (JDK-8220220)
- On AIX, we can read from zero page (address 0) and therefore the test must use a different unreadable address
---
After discussing this, we thing that
- test should use TEST_VM
- the alignment problems are real issues and should be fixed in os::print_hex_dump();
---
Update: for the record, we had the following issues:
- Test did not use TEST_VM so signal handlers could be missing depending on test order
- os::print_hex_dump() had alignment problems.
- On Windows, the gtest will not work since SEH frame is missing. This is a general problem for gtests and I opened a separate issue for this (
- On AIX, we can read from zero page (address 0) and therefore the test must use a different unreadable address
- backported by
-
JDK-8220365 [TESTBUG] Fix crashes in new test hotspot gtest "test_print_hex_dump"
- Resolved
-
JDK-8220433 [TESTBUG] Fix crashes in new test hotspot gtest "test_print_hex_dump"
- Resolved
-
JDK-8220654 [TESTBUG] Fix crashes in new test hotspot gtest "test_print_hex_dump"
- Resolved
-
JDK-8226584 [TESTBUG] Fix crashes in new test hotspot gtest "test_print_hex_dump"
- Resolved
- relates to
-
JDK-8217994 os::print_hex_dump should be more resilient against unreadable memory
- Resolved