8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows #10822

coleenp · 2022-10-21T18:19:36Z

I added another semaphore so the test will run in the order I want it to. Alternate suggestions or corrections welcome.
Running with GHA, and our tier1 in progress.

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows

Reviewers

Patricio Chilano Mateo (@pchilano - Reviewer) ⚠️ Review applies to 6c4eced2
David Holmes (@dholmes-ora - Reviewer) ⚠️ Review applies to 8278537a

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk pull/10822/head:pull/10822
$ git checkout pull/10822

Update a local copy of the PR:
$ git checkout pull/10822
$ git pull https://git.openjdk.org/jdk pull/10822/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 10822

View PR using the GUI difftool:
$ git pr show -t 10822

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/10822.diff

…rmittently on Windows

bridgekeeper · 2022-10-21T18:21:53Z

👋 Welcome back coleenp! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2022-10-21T18:24:19Z

@coleenp The following label will be automatically applied to this pull request:

hotspot-runtime

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

mlbridge · 2022-10-21T18:27:56Z

Webrevs

pchilano

Looks good to me. But I think mainSync is not needed now. Thread 1 will always get to wait() before Thread 2 reaches the notify() call while loading D, so that should be enough to avoid the lost notification issue.

openjdk · 2022-10-21T20:46:15Z

@coleenp This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows

Reviewed-by: pchilanomate, dholmes

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 59 new commits pushed to the master branch:

46e6aee: 8295823: Use enhanced-for cycle instead of Enumeration in java.naming
8e5d680: 8293093: NPE in P11KeyStore.getID
3135914: 8265441: IGV: select block nodes by clicking on it
78454b6: 8295879: JFR: Remove warnings in the jdk.jfr module
33ff80d: 8295402: G1: Tighten assertion in G1ScanHRForRegionClosure::scan_heap_roots
303548b: 8294752: G1: Remove redundant checks in check_obj_during_refinement
cf121df: 8295889: NMT preinit code does not handle allocation errors
772be2e: 8295868: 32-bit Windows build failures after JDK-8294466
e173fd7: 8295885: GHA: Bump gcc versions
b2878c9: 8295710: remove os::dll_file_extension
... and 49 more: https://git.openjdk.org/jdk/compare/0c13d66622a8c2be654bb867aa8c17421d1557ca...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

coleenp · 2022-10-21T23:04:17Z

Thanks Patricio, I don't need mainSync anymore because the first thread will wait.

dholmes-ora

Your are doing an unconditional wait() which in theory could be affected by a spurious wakeup. If you actually do:

while (!DisLoading) {
  wait();
}

with a corresponding

DisLoading = true;
notify();

then I think you could dispense with the semaphore altogether.

…loader locked region, and guard against spuriouss wakeups.

coleenp · 2022-10-24T21:28:32Z

I need to both handle 1. the spurious wakeup from wait(), and 2. the case where the second thread loads everything, and then the first thread then waits, after D's notification - the missed notification.

dholmes-ora · 2022-10-24T22:26:36Z

I need to both handle 1. the spurious wakeup from wait(), and 2. the case where the second thread loads everything, and then the first thread then waits, after D's notification - the missed notification.

That is what DIsLoaded indicates. There's no missed notification because you won't wait if D is already loaded.

coleenp · 2022-10-25T12:13:44Z

It won't test what I'm trying to test though, but maybe it's ok. I added the test for illustrative purposes.

dholmes-ora

Sorry Coleen some of the intermediate changes here confused me. The original use of mainSync is slowing down thread1 (loading A) so that it can't proceed until thread 2 starts to load C and has the lock of L2 (which sets up the potential deadlock scenario). That could be added back in conjunction with the dIsLoaded usage. In both cases thread 2 could always get to go first and load C and D before thread 1 has a chance to do anything.

dholmes-ora · 2022-10-25T12:51:27Z

test/hotspot/jtreg/runtime/ParallelLoad/SuperWait/SuperWaitTest.java

@@ -36,19 +36,17 @@

 public class SuperWaitTest {

-    private static Semaphore mainSync = null;
+    private static volatile boolean dIsLoading = false;


This doesn't need to be volatile as it is only accessed under the classloader lock.

coleenp · 2022-10-25T13:15:52Z

My n-1 change makes thread 2 wait for loading C and D until thread1 is in the locked region for loading A (which was the deadlock scenario before the wait() is introduced). It doesn't really matter if thread 2 goes first with the dIsLoading change. Either way, the test will complete and not miss a notify.

dholmes-ora

Okay this seems fine from a logic perspective. There's one nit below that I missed before but up to you whether to change it.
Thanks.

dholmes-ora · 2022-10-26T00:30:25Z

test/hotspot/jtreg/runtime/ParallelLoad/SuperWait/SuperWaitTest.java

@@ -36,19 +36,17 @@

 public class SuperWaitTest {

-    private static Semaphore mainSync = null;
+    private static boolean dIsLoading = false;


Nit: sorry just noticed the placement here. This is really an instance field of MyLoaderOne.

I fixed it.

coleenp · 2022-10-26T17:26:45Z

Thanks for the reviews Patricio and David.
/integrate

openjdk · 2022-10-26T17:28:51Z

Going to push as commit a8450b3.
Since your change was applied there have been 59 commits pushed to the master branch:

46e6aee: 8295823: Use enhanced-for cycle instead of Enumeration in java.naming
8e5d680: 8293093: NPE in P11KeyStore.getID
3135914: 8265441: IGV: select block nodes by clicking on it
78454b6: 8295879: JFR: Remove warnings in the jdk.jfr module
33ff80d: 8295402: G1: Tighten assertion in G1ScanHRForRegionClosure::scan_heap_roots
303548b: 8294752: G1: Remove redundant checks in check_obj_during_refinement
cf121df: 8295889: NMT preinit code does not handle allocation errors
772be2e: 8295868: 32-bit Windows build failures after JDK-8294466
e173fd7: 8295885: GHA: Bump gcc versions
b2878c9: 8295710: remove os::dll_file_extension
... and 49 more: https://git.openjdk.org/jdk/compare/0c13d66622a8c2be654bb867aa8c17421d1557ca...master

Your commit was automatically rebased without conflicts.

openjdk · 2022-10-26T17:29:20Z

@coleenp Pushed as commit a8450b3.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails inte…

6c4eced

…rmittently on Windows

openjdk bot added the rfr Pull request is ready for review label Oct 21, 2022

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Oct 21, 2022

pchilano approved these changes Oct 21, 2022

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Oct 21, 2022

Thanks Patricio

d0089bb

dholmes-ora reviewed Oct 24, 2022

View reviewed changes

Fix the logic some more. Wait for thread2 without being in the class …

e24117f

…loader locked region, and guard against spuriouss wakeups.

David's fix - allow thread 2 to run first.

b2decdf

dholmes-ora reviewed Oct 25, 2022

View reviewed changes

remove volatile

8278537

dholmes-ora approved these changes Oct 26, 2022

View reviewed changes

Fix nit.

aad3a70

openjdk bot added the integrated Pull request has been integrated label Oct 26, 2022

openjdk bot closed this Oct 26, 2022

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Oct 26, 2022

coleenp deleted the loading-tests branch October 26, 2022 18:03

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows #10822

8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows #10822

coleenp commented Oct 21, 2022 •

edited by openjdk bot

Loading

bridgekeeper bot commented Oct 21, 2022

openjdk bot commented Oct 21, 2022

mlbridge bot commented Oct 21, 2022 •

edited

Loading

pchilano left a comment

openjdk bot commented Oct 21, 2022 •

edited

Loading

coleenp commented Oct 21, 2022

dholmes-ora left a comment •

edited

Loading

coleenp commented Oct 24, 2022 •

edited

Loading

dholmes-ora commented Oct 24, 2022

coleenp commented Oct 25, 2022

dholmes-ora left a comment

dholmes-ora Oct 25, 2022

coleenp commented Oct 25, 2022

dholmes-ora left a comment

dholmes-ora Oct 26, 2022

coleenp Oct 26, 2022

coleenp commented Oct 26, 2022

openjdk bot commented Oct 26, 2022

openjdk bot commented Oct 26, 2022

8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows #10822

8295713: runtime/ParallelLoad/SuperWait/SuperWaitTest.java fails intermittently on Windows #10822

Conversation

coleenp commented Oct 21, 2022 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Reviewing

bridgekeeper bot commented Oct 21, 2022

openjdk bot commented Oct 21, 2022

mlbridge bot commented Oct 21, 2022 • edited Loading

Webrevs

pchilano left a comment

Choose a reason for hiding this comment

openjdk bot commented Oct 21, 2022 • edited Loading

coleenp commented Oct 21, 2022

dholmes-ora left a comment • edited Loading

Choose a reason for hiding this comment

coleenp commented Oct 24, 2022 • edited Loading

dholmes-ora commented Oct 24, 2022

coleenp commented Oct 25, 2022

dholmes-ora left a comment

Choose a reason for hiding this comment

dholmes-ora Oct 25, 2022

Choose a reason for hiding this comment

coleenp commented Oct 25, 2022

dholmes-ora left a comment

Choose a reason for hiding this comment

dholmes-ora Oct 26, 2022

Choose a reason for hiding this comment

coleenp Oct 26, 2022

Choose a reason for hiding this comment

coleenp commented Oct 26, 2022

openjdk bot commented Oct 26, 2022

openjdk bot commented Oct 26, 2022

coleenp commented Oct 21, 2022 •

edited by openjdk bot

Loading

mlbridge bot commented Oct 21, 2022 •

edited

Loading

openjdk bot commented Oct 21, 2022 •

edited

Loading

dholmes-ora left a comment •

edited

Loading

coleenp commented Oct 24, 2022 •

edited

Loading