8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable #13634

shipilev · 2023-04-25T09:01:13Z

As seen in some CI runs, doing Object.wait(1) to inflate 10K monitors can take quite a while. One of the test cases implicitly relies on threshold heuristics kicking in before the guaranteed interval of 5 seconds expires. Even if we wait for 1ms for each wait call, we only inflate 5K monitors in those 5 seconds. Which is okay for 5s and 10% threshold test, which would fire after about 1K monitors inflated. But we can also stall for longer -- because wait(1) is not guaranteed to return in 1ms, subject to locking granularity -- which means we can inflate less that 1K monitors, which would mean the threshold heuristics would not fire.

This fix strengthens the test in two ways:

Uses multiple threads to inflate the monitors, so that we are able to eat the cost of wait(1).
Drops the threshold heuristics threshold to minimum value to increase the chances of it firing.

This also makes the test faster.

@dcubed-ojdk, you might want to run this through your CI to see if it fixes the test troubles.

Additional testing:

Ad-hoc runs of the test, 100 iterations without trouble.

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-8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable

Reviewers

Thomas Stuefe (@tstuefe - Reviewer) ⚠️ Review applies to 76fa7114
Daniel D. Daugherty (@dcubed-ojdk - Reviewer) ⚠️ Review applies to 76fa7114

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/13634/head:pull/13634
$ git checkout pull/13634

Update a local copy of the PR:
$ git checkout pull/13634
$ git pull https://git.openjdk.org/jdk.git pull/13634/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 13634

View PR using the GUI difftool:
$ git pr show -t 13634

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/13634.diff

Webrev

Link to Webrev Comment

bridgekeeper · 2023-04-25T09:01:56Z

👋 Welcome back shade! 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 · 2023-04-25T09:04:17Z

@shipilev 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 · 2023-04-25T09:08:14Z

Webrevs

tstuefe

LGTM. Why did you steer away from the i-hash solution though, I thought that was rather neat.

openjdk · 2023-04-25T16:15:04Z

@shipilev 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:

8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable

Reviewed-by: stuefe, dcubed

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 12 new commits pushed to the master branch:

44d9f55: 8306072: Open source several AWT MouseInfo related tests
cc894d8: 8303466: C2: failed: malformed control flow. Limit type made precise with MaxL/MinL
ed1ebd2: 8306652: Open source AWT MenuItem related tests
f3e8bd1: 8306755: Open source few Swing JComponent and AbstractButton tests
1c1a73f: 8302908: RISC-V: Support masked vector arithmetic instructions for Vector API
adf62fe: 8304918: Remove unused decl field from AnnotatedType implementations
00b1eac: 8306031: Update IANA Language Subtag Registry to Version 2023-04-13
88d9ebf: 8306752: Open source several container and component AWT tests
1c2dadc: 8306683: Open source several clipboard and color AWT tests
b372f28: 8306753: Open source several container AWT tests
... and 2 more: https://git.openjdk.org/jdk/compare/28829f308fe6314388c9a47b91273bcf81eb806c...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.

shipilev · 2023-04-25T16:16:54Z

LGTM. Why did you steer away from the i-hash solution though, I thought that was rather neat.

Previous version of this PR used the take-IHC-while-locked trick to inflate the monitors, but I figured it would not work after JDK-8291555: Implement alternative fast-locking scheme is here, and would quite probably be even less reliable as Lilliput progresses towards off-side IHC table.

dcubed-ojdk · 2023-04-25T20:38:17Z

LGTM. Why did you steer away from the i-hash solution though, I thought that was rather neat.

Previous version of this PR used the take-IHC-while-locked trick to inflate the monitors, but I figured it would not work after JDK-8291555: Implement alternative fast-locking scheme is here, and would quite probably be even less reliable as Lilliput progresses towards off-side IHC table.

Hmmm... I don't remember seeing anything in JDK-8291555 that would break the
take-IHC-while-locked trick. I use that trick in some of my monitor stress tests...
It would be sad to see that go away...

dcubed-ojdk

Looks like a reasonable fix so thumbs up.
I'm testing in Mach5 now.

test/hotspot/jtreg/runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java

dcubed-ojdk · 2023-04-25T21:52:02Z

Mach5 Tier1 testing passed. It doesn't look like the test is
executed in Tier[23].

shipilev · 2023-04-26T07:58:44Z

Previous version of this PR used the take-IHC-while-locked trick to inflate the monitors, but I figured it would not work after JDK-8291555: Implement alternative fast-locking scheme is here, and would quite probably be even less reliable as Lilliput progresses towards off-side IHC table.

Hmmm... I don't remember seeing anything in JDK-8291555 that would break the take-IHC-while-locked trick. I use that trick in some of my monitor stress tests... It would be sad to see that go away...

Perhaps I was too cautious about this. I figured that the practical reason to inflate the stack-locked object when IHC is requested is to deal with various races about the IHC installation in the displaced header. And since JDK-8291555 changes the stack-locking to not replace the header fully (only changing a bit, AFAIU), then IHC installation becomes much easier. But after reading the description again, I think JDK-8291555 only takes care of IHC on fastpath, i.e. when it is already installed. So take-IHC-while-locked might still work.

Still, I don't feel quite safe with relying on it, because it seems (famous last words) an easy future change to allow IHC installation into stack-locked object header after JDK-8291555, and I think there is a tentative plan to move IHC out of the header completely with 32-bit headers, which would make inflation even less of the practical behavior. So I believe wait(1) remains the only future-proof way to inflate the monitors, before we start to mess with JVM settings.

shipilev · 2023-04-26T07:59:00Z

Mach5 Tier1 testing passed. It doesn't look like the test is executed in Tier[23].

Thank you! I'll re-run some local testing and then integrate.

shipilev · 2023-04-26T11:09:46Z

Local testing looks good. GHAs look good. I am integrating now.

/integrate

openjdk · 2023-04-26T11:12:04Z

Going to push as commit 9ad6dc8.
Since your change was applied there have been 16 commits pushed to the master branch:

c5910fa: 8306749: Make CardTable::invalidate non-virtual
86f41a4: 8306735: G1: G1FullGCScope remove unnecessary member _explicit_gc
d747698: 8306823: Native memory leak in SharedRuntime::notify_jvmti_unmount/mount.
8d89992: 8298189: Regression in SPECjvm2008-MonteCarlo for pre-Cascade Lake Intel processors
44d9f55: 8306072: Open source several AWT MouseInfo related tests
cc894d8: 8303466: C2: failed: malformed control flow. Limit type made precise with MaxL/MinL
ed1ebd2: 8306652: Open source AWT MenuItem related tests
f3e8bd1: 8306755: Open source few Swing JComponent and AbstractButton tests
1c1a73f: 8302908: RISC-V: Support masked vector arithmetic instructions for Vector API
adf62fe: 8304918: Remove unused decl field from AnnotatedType implementations
... and 6 more: https://git.openjdk.org/jdk/compare/28829f308fe6314388c9a47b91273bcf81eb806c...master

Your commit was automatically rebased without conflicts.

openjdk · 2023-04-26T11:12:18Z

@shipilev Pushed as commit 9ad6dc8.

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

Aleksei Shipilev added 2 commits April 25, 2023 10:43

Fix

06a3cb5

Drop MUDT to 1 to increase the chances

1c259ab

openjdk bot added the rfr Pull request is ready for review label Apr 25, 2023

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Apr 25, 2023

Parallel inflation instead of relying on hashCode

76fa711

tstuefe approved these changes Apr 25, 2023

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Apr 25, 2023

dcubed-ojdk approved these changes Apr 25, 2023

View reviewed changes

test/hotspot/jtreg/runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java Outdated Show resolved Hide resolved

Aleksei Shipilev added 2 commits April 26, 2023 10:03

Drop redundant field

c7b507e

Merge branch 'master' into JDK-8306774-async-test

044c835

openjdk bot added the integrated Pull request has been integrated label Apr 26, 2023

openjdk bot closed this Apr 26, 2023

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Apr 26, 2023

shipilev deleted the JDK-8306774-async-test branch May 3, 2023 09:50

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable #13634

8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable #13634

shipilev commented Apr 25, 2023 •

edited by openjdk bot

Loading

bridgekeeper bot commented Apr 25, 2023

openjdk bot commented Apr 25, 2023

mlbridge bot commented Apr 25, 2023 •

edited

Loading

tstuefe left a comment

openjdk bot commented Apr 25, 2023 •

edited

Loading

shipilev commented Apr 25, 2023

dcubed-ojdk commented Apr 25, 2023

dcubed-ojdk left a comment

dcubed-ojdk commented Apr 25, 2023

shipilev commented Apr 26, 2023

shipilev commented Apr 26, 2023

shipilev commented Apr 26, 2023

openjdk bot commented Apr 26, 2023

openjdk bot commented Apr 26, 2023

8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable #13634

8306774: Make runtime/Monitor/GuaranteedAsyncDeflationIntervalTest.java more reliable #13634

Conversation

shipilev commented Apr 25, 2023 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Reviewing

Webrev

bridgekeeper bot commented Apr 25, 2023

openjdk bot commented Apr 25, 2023

mlbridge bot commented Apr 25, 2023 • edited Loading

Webrevs

tstuefe left a comment

Choose a reason for hiding this comment

openjdk bot commented Apr 25, 2023 • edited Loading

shipilev commented Apr 25, 2023

dcubed-ojdk commented Apr 25, 2023

dcubed-ojdk left a comment

Choose a reason for hiding this comment

dcubed-ojdk commented Apr 25, 2023

shipilev commented Apr 26, 2023

shipilev commented Apr 26, 2023

shipilev commented Apr 26, 2023

openjdk bot commented Apr 26, 2023

openjdk bot commented Apr 26, 2023

shipilev commented Apr 25, 2023 •

edited by openjdk bot

Loading

mlbridge bot commented Apr 25, 2023 •

edited

Loading

openjdk bot commented Apr 25, 2023 •

edited

Loading