Question on backporting JDK-8169739

[johnsjiang(江莎)]

Hi Severin,Thanks for your reply!


> May I ask why? Is this a critical backport?


We encountered the same problem as JDK-8171051, so originally I just wanted to backport this issue.
However, I found the fix for JDK-8171051 modified not only LinkedBlockingQueue, but LinkedBlockingDeque as well.
Then I supposed it may need to backport JDK-8169739 at first.


> OpenJDK 8u is in long term maintenance mode.


JDK-8169739 is a bug, but not an enhancement, so I thought it possibly has a chance to fix this bug for OpenJDK 8u.


> I'd suggest to upgrade the JDK if your users need the fix.


Unfortunately, it's not easy to upgrade JDK from 8 to 11+ for this existing system at the moment.


> That seems a large too risky change to backport as-is from JDK 9. If
> you can find a minimal version of it that fixes the problem, get
> sufficient review(s) from domain experts, perform some extensive
> testing, we might be able to discuss again.


I really hesitated to backport such a big patch, so I sent my last mail to this alias to discuss it.
I supposed only LinkedBlockingQueue related changes should be backported.


With a bit more investigation, I found the associated tck tests were introduced by another JBS issue, exactly JDK-8146467.
Hmm, the deal is becoming bigger.


Best regards,
John Jiang




                    Severin Gehwolf<sgehwolf at redhat.com> 在 2024年2月28日 周三 18:52 写道：

On Wed, 2024-02-28 at 14:49 +0800, johnsjiang(江莎) wrote:
> Hi,
> I want to backport JDK-8169739 [1] to OpenJDK 8u

May I ask why? Is this a critical backport? OpenJDK 8u is in long term
maintenance mode. I'd suggest to upgrade the JDK if your users need the
fix.

> but find the associated
> commit [2] contains the changes for multiple issues, including JDK-8143577,
> JDK-8169679, JDK-8167202, JDK-8164793, JDK-8169739 and JDK-8169738.
> 
> Do I have to backport all the issues in a batch?
> Any suggestion?

That seems a large too risky change to backport as-is from JDK 9. If
you can find a minimal version of it that fixes the problem, get
sufficient review(s) from domain experts, perform some extensive
testing, we might be able to discuss again.

Thanks,
Severin

> [1] https://bugs.openjdk.org/browse/JDK-8169739
> [2] https://github.com/openjdk/jdk/commit/3d0d86185e189b84e4ee2802314e1596ad66a779
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/jdk8u-dev/attachments/20240229/ea096ba5/attachment-0001.htm>