On Wed, 12 Jul 2023 08:01:55 GMT, Serguei Spitsyn <[email protected]> wrote:
>> This is an issue with a dynamic load of a JVMTI agent into running VM.
>> The `VM_SetNotifyJvmtiEventsMode` enabling operation makes a call to the
>> function `count_transitions_and_correct_jvmti_thread_states()`. This
>> function in its turn make a call to the function
>> `correct_jvmti_thread_state()`. But it does it conditionally, only if the
>> field `_whitebox_used` is `true`.
>> The test provided in the bug report showed that it has to be called
>> unconditionally as the assumption that it is only needed on the subsequent
>> `notifyJvmti` enabling is incorrect.
>>
>> Then the field `_whitebox_used` is not needed anymore and removed in this
>> fix.
>> Some obsolete comments are removed or updated.
>>
>> New test is added:
>> `test/hotspot/jtreg/serviceability/jvmti/vthread/VThreadTLSTest`.
>> It is failed without the fix and passed with the fix.
>>
>> Testing:
>> - ran new test
>> `test/hotspot/jtreg/serviceability/jvmti/vthread/VThreadTLSTest`
>> - mach5 tiers 1-5 are good
>
> Serguei Spitsyn has updated the pull request incrementally with one
> additional commit since the last revision:
>
> review: removed unneeded @compile commands from new test
Looks good other than one minor typo.
test/hotspot/jtreg/serviceability/jvmti/vthread/VThreadTLSTest/VThreadTLSTest.java
line 68:
> 66: return;
> 67: }
> 68: for (int repetion = 0; repetion < 10; repetion++) {
repetion -> repetition
-------------
Marked as reviewed by cjplummer (Reviewer).
PR Review: https://git.openjdk.org/jdk/pull/14842#pullrequestreview-1527411302
PR Review Comment: https://git.openjdk.org/jdk/pull/14842#discussion_r1261855278
