8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved #18417

limingliu-ampere · 2024-03-21T02:49:48Z

The testcase failed on Oracle CI since JDK-8315923. The root cause is that Oracle CI runs Linux-5.4.17-UEK where the value of MADV_POPULATE_WRITE (23) is used as MADV_DONTEXEC which is not supported by upstream. This PR solves the testcase failure by checking the support of (MADV_POPULATE_WRITE_value + 1) for UEK releases.

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

Issues

JDK-8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved (Bug - P2)
JDK-8325218: gc/parallel/TestAlwaysPreTouchBehavior.java fails (Bug - P3)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/18417/head:pull/18417
$ git checkout pull/18417

Update a local copy of the PR:
$ git checkout pull/18417
$ git pull https://git.openjdk.org/jdk.git pull/18417/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 18417

View PR using the GUI difftool:
$ git pr show -t 18417

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/18417.diff

Webrev

Link to Webrev Comment

bridgekeeper · 2024-03-21T02:50:16Z

👋 Welcome back limingliu-ampere! 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 · 2024-03-21T02:50:22Z

❗ This change is not yet ready to be integrated.
See the Progress checklist in the description for automated requirements.

limingliu-ampere · 2024-03-21T02:50:35Z

/issue add 8325218

openjdk · 2024-03-21T02:50:46Z

@limingliu-ampere
Adding additional issue to issue list: 8325218: gc/parallel/TestAlwaysPreTouchBehavior.java fails.

openjdk · 2024-03-21T02:51:27Z

@limingliu-ampere 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.

…ccordingly

openjdk · 2024-03-21T02:56:18Z

@limingliu-ampere Please do not rebase or force-push to an active PR as it invalidates existing review comments. Note for future reference, the bots always squash all changes into a single commit automatically as part of the integration. See OpenJDK Developers’ Guide for more information.

mlbridge · 2024-03-21T02:59:38Z

Webrevs

01: Full - Incremental (005259f5)
00: Full (192855f8)

dholmes-ora · 2024-03-21T05:16:56Z

src/hotspot/os/linux/os_linux.cpp

+    // See https://github.com/oracle/linux-uek/issues/23
+    const int flag = MADV_POPULATE_WRITE + (os::Linux::_is_uek_release ? 1 : 0);
+    FLAG_SET_DEFAULT(UseMadvPopulateWrite, (::madvise(0, 0, flag) == 0));


I read the supporting material and I still can't get this sorted in my head. From my reading some UEK versions have this problem and some do not. But you do this "trick" for all UEK versions and I just don't understand how that works. ??

Personally, I think it is safe to check 24 for all UEK versions:

for versions prior to 5.15 UEK, 23 is supported as MADV_DONTEXEC, while 24 is not supported;

for 5.15 (, 5.16 and 5.17 if existed) UEK, 23 is supported as MADV_POPULATE_WRITE, while 24 is supported as MADV_DOEXEC;

for 5.18 UEK or newer, both 23 and 24 are supported as the same functions as upstream.

So what is the affect of using MADV_DOEXEC when the intent was to use MADV_POPULATE_WRITE?

MADV_DOEXEC is not used in os::pd_pretouch_memory, but just used in os::init_2 for 5.15 UEK. The key is MADV_POPULATE_WRITE is supported as 23 on UEK releases if and only if 24 is supported. If the code is confusing, I will check whether the version of kernel is newer than 5.14 here rather than checking the support of 24.

limingliu-ampere · 2024-04-03T07:51:33Z

/withdraw

openjdk · 2024-04-03T07:51:51Z

@limingliu-ampere Unknown command withdraw - for a list of valid commands use /help.

limingliu-ampere · 2024-04-03T07:52:13Z

/help

openjdk · 2024-04-03T07:52:51Z

@limingliu-ampere Available commands:

approval - request for maintainer's approval
approve - null
author - sets an overriding author to be used in the commit when the PR is integrated
backport - create a backport
cc - add or remove an additional classification label
clean - Mark the backport pull request as a clean backport
contributor - adds or removes additional contributors for a PR
covered - used when employer has signed the OCA
csr - require a compatibility and specification request (CSR) for this pull request
help - shows this text
integrate - performs integration of the changes in the PR
issue - edit the list of issues that this PR solves
jep - require a JDK Enhancement Proposal (JEP) for this pull request
label - add or remove an additional classification label
open - Set the pull request state to "open"
reviewer - manage additional reviewers for a PR
reviewers - set the number of additional required reviewers for this PR
signed - used after signing the OCA
solves - edit the list of issues that this PR solves
sponsor - performs integration of a PR that is authored by a non-committer
summary - updates the summary in the commit message
test - used to run tests

tstuefe · 2024-05-22T06:57:12Z

Please reopen this, and just run the test with -XX:-UseMadvPopulateWrite. That should solve the problem for the test, no?

The test's purpose is not to test Oracle Linux' inability to handle madvise flags correctly but to test -XX:+AlwaysPreTouchStacks.

We can and should deal with the Oracle Linux problem independently (Note that I reported the problem to Oracle secalert back in March, and they promised a kernel fix).

stefank · 2024-05-22T07:45:36Z

Please reopen this, and just run the test with -XX:-UseMadvPopulateWrite. That should solve the problem for the test, no?

The test's purpose is not to test Oracle Linux' inability to handle madvise flags correctly but to test -XX:+AlwaysPreTouchStacks.

We can and should deal with the Oracle Linux problem independently (Note that I reported the problem to Oracle secalert back in March, and they promised a kernel fix).

This is being dealt with in #18592, right? (Which I now see needs to be sponsored)

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Mar 21, 2024

Detect UEK releases and adjust the detection of MADV_POPULATE_WRITE a…

192855f

…ccordingly

limingliu-ampere force-pushed the 8324781 branch from 7dbc1df to 192855f Compare March 21, 2024 02:55

openjdk bot added the rfr Pull request is ready for review label Mar 21, 2024

dholmes-ora reviewed Mar 21, 2024

View reviewed changes

Remove 8325218 from the problem list

005259f

limingliu-ampere closed this Apr 3, 2024

tstuefe mentioned this pull request May 22, 2024

8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved #18592

Closed

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved #18417

8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved #18417

limingliu-ampere commented Mar 21, 2024 •

edited by openjdk bot

Loading

bridgekeeper bot commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

limingliu-ampere commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

mlbridge bot commented Mar 21, 2024 •

edited

Loading

dholmes-ora Mar 21, 2024

limingliu-ampere Mar 21, 2024 •

edited

Loading

dholmes-ora Mar 25, 2024

limingliu-ampere Mar 25, 2024

limingliu-ampere commented Apr 3, 2024

openjdk bot commented Apr 3, 2024

limingliu-ampere commented Apr 3, 2024

openjdk bot commented Apr 3, 2024

tstuefe commented May 22, 2024

stefank commented May 22, 2024

8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved #18417

8324781: runtime/Thread/TestAlwaysPreTouchStacks.java failed with Expected a higher ratio between stack committed and reserved #18417

Conversation

limingliu-ampere commented Mar 21, 2024 • edited by openjdk bot Loading

Progress

Issues

Reviewing

Webrev

bridgekeeper bot commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

limingliu-ampere commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

openjdk bot commented Mar 21, 2024

mlbridge bot commented Mar 21, 2024 • edited Loading

Webrevs

dholmes-ora Mar 21, 2024

Choose a reason for hiding this comment

limingliu-ampere Mar 21, 2024 • edited Loading

Choose a reason for hiding this comment

dholmes-ora Mar 25, 2024

Choose a reason for hiding this comment

limingliu-ampere Mar 25, 2024

Choose a reason for hiding this comment

limingliu-ampere commented Apr 3, 2024

openjdk bot commented Apr 3, 2024

limingliu-ampere commented Apr 3, 2024

openjdk bot commented Apr 3, 2024

tstuefe commented May 22, 2024

stefank commented May 22, 2024

limingliu-ampere commented Mar 21, 2024 •

edited by openjdk bot

Loading

mlbridge bot commented Mar 21, 2024 •

edited

Loading

limingliu-ampere Mar 21, 2024 •

edited

Loading