JDK-8275704: Metaspace::contains() should be threadsafe #6060

tstuefe · 2021-10-21T09:14:26Z

Metaspace::contains() is used in many places. It is not threadsafe since it walks the list of metaspace mappings, which can be altered concurrently. This is suspected to be the cause of JDK-8271124.

Currently, it does not lock, and adding a lock is not realistic either. It should work lockless.

This patch builds atop of https://bugs.openjdk.java.net/browse/JDK-8275582, which removed the old (pre JEP 387) technique of uncommitting metaspace memory. As a side effect, that patch changed the mapping list to an add-only structure. The only remaining place where it gets modified is in VirtualSpaceList::create_node(). Modifications are synchronized via lock. The only place where we walk the list locklessly is in Metaspace::contains(). This patch adds the appropriate memory barriers to those two places.

Tests:

GHAs
SAP nightlies (queued)

Progress

Change must not contain extraneous whitespace
Commit message must refer to an issue
Change must be properly reviewed

Issue

JDK-8275704: Metaspace::contains() should be threadsafe

Reviewers

Coleen Phillimore (@coleenp - Reviewer) ⚠️ Review applies to bfe54c1
David Holmes (@dholmes-ora - Reviewer) ⚠️ Review applies to bfe54c1

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk pull/6060/head:pull/6060
$ git checkout pull/6060

Update a local copy of the PR:
$ git checkout pull/6060
$ git pull https://git.openjdk.java.net/jdk pull/6060/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 6060

View PR using the GUI difftool:
$ git pr show -t 6060

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk/pull/6060.diff

bridgekeeper · 2021-10-21T09:14:45Z

👋 Welcome back stuefe! 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 · 2021-10-21T09:17:22Z

@tstuefe 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 · 2021-10-27T09:12:26Z

Webrevs

01: Full - Incremental (8f6e91fc)
00: Full (bfe54c16)

coleenp

This looks good! Thank you for fixing this.

openjdk · 2021-10-27T12:48:33Z

@tstuefe 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:

8275704: Metaspace::contains() should be threadsafe

Reviewed-by: coleenp, 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 12 new commits pushed to the master branch:

9a3e954: 8274879: Replace uses of StringBuffer with StringBuilder within java.base classes
e6fa5fa: 8276063: ProblemList gtest dll_address_to_function_and_library_name on macosx-generic
809488b: 8276046: codestrings.validate_vm gtest fails on ppc64, s390
93be099: 4718400: Many quantities are held as signed that should be unsigned
168081e: 8270490: Charset.forName() taking fallback default value
a292733: 8275975: Remove dead code in ciInstanceKlass
4060602: 8275800: Redefinition leaks MethodData::_extra_data_lock
485d658: 8275851: Deproblemlist open/test/jdk/javax/swing/JComponent/6683775/bug6683775.java
b3f45f8: 8275689: [TESTBUG] Use color tolerance only for XRender in BlitRotateClippedArea test
6c05cc9: 8275863: Use encodeASCII for ASCII-compatible DoubleByte encodings
... and 2 more: https://git.openjdk.java.net/jdk/compare/9f75d5ce500886b32175cc541939b7f0eee190ca...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.

tstuefe · 2021-10-27T12:55:16Z

This looks good! Thank you for fixing this.

Thank you Coleen!

zhengyu123 · 2021-10-27T13:07:25Z

src/hotspot/share/memory/metaspace/virtualSpaceList.cpp

@@ -186,7 +189,8 @@ void VirtualSpaceList::verify() const {

 // Returns true if this pointer is contained in one of our nodes.
 bool VirtualSpaceList::contains(const MetaWord* p) const {
-  const VirtualSpaceNode* vsn = _first_node;
+  // Note: needs to work without locks.
+  const VirtualSpaceNode* vsn = Atomic::load_acquire(&_first_node);


I believe vsn->next() now also requires load_acquire semantics.

The internal nodes are already protected by the acquire/release semantics used for accessing _first_node, which is the only mutable field.

That's what I figured too.

Also, if that were wrong we would have the same problem in CLDG. Where, in contrast to this list nodes get deleted too (see ClassLoaderDataGraph::clear_claimed_marks()).

dholmes-ora

Hi Thomas,

This seems fine to me.

Thanks,
David

dholmes-ora · 2021-10-28T02:50:37Z

src/hotspot/share/memory/metaspace/virtualSpaceList.cpp

@@ -186,7 +189,8 @@ void VirtualSpaceList::verify() const {

 // Returns true if this pointer is contained in one of our nodes.
 bool VirtualSpaceList::contains(const MetaWord* p) const {
-  const VirtualSpaceNode* vsn = _first_node;
+  // Note: needs to work without locks.
+  const VirtualSpaceNode* vsn = Atomic::load_acquire(&_first_node);


The internal nodes are already protected by the acquire/release semantics used for accessing _first_node, which is the only mutable field.

dholmes-ora · 2021-10-28T02:51:15Z

src/hotspot/share/memory/metaspace/virtualSpaceList.hpp

+//  never to be destroyed.
+//
+// The list will only be modified under lock protection, but may be
+//  read asynchronously without lock.


asynchronously? Did you mean concurrently?

Right, I'll correct it before pushing.

tstuefe · 2021-10-28T05:28:55Z

Thanks @coleenp, @zhengyu123, @dholmes-ora!

Note that one of the next changes will be an increase in size of the individual mappings. Today they are two root chunks sized (8MB) and by making them bigger we will have fewer of them, which will make traversing the list in contains() faster.

/integrate

openjdk · 2021-10-28T05:30:07Z

Going to push as commit d9b0138.
Since your change was applied there have been 12 commits pushed to the master branch:

9a3e954: 8274879: Replace uses of StringBuffer with StringBuilder within java.base classes
e6fa5fa: 8276063: ProblemList gtest dll_address_to_function_and_library_name on macosx-generic
809488b: 8276046: codestrings.validate_vm gtest fails on ppc64, s390
93be099: 4718400: Many quantities are held as signed that should be unsigned
168081e: 8270490: Charset.forName() taking fallback default value
a292733: 8275975: Remove dead code in ciInstanceKlass
4060602: 8275800: Redefinition leaks MethodData::_extra_data_lock
485d658: 8275851: Deproblemlist open/test/jdk/javax/swing/JComponent/6683775/bug6683775.java
b3f45f8: 8275689: [TESTBUG] Use color tolerance only for XRender in BlitRotateClippedArea test
6c05cc9: 8275863: Use encodeASCII for ASCII-compatible DoubleByte encodings
... and 2 more: https://git.openjdk.java.net/jdk/compare/9f75d5ce500886b32175cc541939b7f0eee190ca...master

Your commit was automatically rebased without conflicts.

openjdk · 2021-10-28T05:30:24Z

@tstuefe Pushed as commit d9b0138.

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

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Oct 21, 2021

tstuefe changed the title ~~JDK-8275704: Metaspace::contains() should the threadsafe~~ JDK-8275704: Metaspace::contains() should be threadsafe Oct 21, 2021

make metaspace contains threadsafe

bfe54c1

tstuefe force-pushed the JDK-8275704-metaspace-contains-should-be-threadsafe branch from 7ddf71e to bfe54c1 Compare October 27, 2021 05:53

tstuefe marked this pull request as ready for review October 27, 2021 09:07

openjdk bot added the rfr Pull request is ready for review label Oct 27, 2021

coleenp approved these changes Oct 27, 2021

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Oct 27, 2021

zhengyu123 suggested changes Oct 27, 2021

View reviewed changes

dholmes-ora approved these changes Oct 28, 2021

View reviewed changes

fix typo

8f6e91f

openjdk bot closed this Oct 28, 2021

openjdk bot added integrated Pull request has been integrated and removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Oct 28, 2021

tstuefe deleted the JDK-8275704-metaspace-contains-should-be-threadsafe branch October 28, 2021 05:36

JDK-8275704: Metaspace::contains() should be threadsafe #6060

JDK-8275704: Metaspace::contains() should be threadsafe #6060

tstuefe commented Oct 21, 2021 •

edited by openjdk bot

bridgekeeper bot commented Oct 21, 2021

openjdk bot commented Oct 21, 2021

mlbridge bot commented Oct 27, 2021 •

edited

coleenp left a comment

openjdk bot commented Oct 27, 2021 •

edited

tstuefe commented Oct 27, 2021

zhengyu123 Oct 27, 2021

dholmes-ora Oct 28, 2021

tstuefe Oct 28, 2021

dholmes-ora left a comment

dholmes-ora Oct 28, 2021

dholmes-ora Oct 28, 2021

tstuefe Oct 28, 2021

tstuefe commented Oct 28, 2021

openjdk bot commented Oct 28, 2021

openjdk bot commented Oct 28, 2021

JDK-8275704: Metaspace::contains() should be threadsafe #6060

JDK-8275704: Metaspace::contains() should be threadsafe #6060

Conversation

tstuefe commented Oct 21, 2021 • edited by openjdk bot

Progress

Issue

Reviewers

Reviewing

bridgekeeper bot commented Oct 21, 2021

openjdk bot commented Oct 21, 2021

mlbridge bot commented Oct 27, 2021 • edited

Webrevs

coleenp left a comment

Choose a reason for hiding this comment

openjdk bot commented Oct 27, 2021 • edited

tstuefe commented Oct 27, 2021

zhengyu123 Oct 27, 2021

Choose a reason for hiding this comment

dholmes-ora Oct 28, 2021

Choose a reason for hiding this comment

tstuefe Oct 28, 2021

Choose a reason for hiding this comment

dholmes-ora left a comment

Choose a reason for hiding this comment

dholmes-ora Oct 28, 2021

Choose a reason for hiding this comment

dholmes-ora Oct 28, 2021

Choose a reason for hiding this comment

tstuefe Oct 28, 2021

Choose a reason for hiding this comment

tstuefe commented Oct 28, 2021

openjdk bot commented Oct 28, 2021

openjdk bot commented Oct 28, 2021

tstuefe commented Oct 21, 2021 •

edited by openjdk bot

mlbridge bot commented Oct 27, 2021 •

edited

openjdk bot commented Oct 27, 2021 •

edited