8276205: Shenandoah: CodeCache_lock should always be held for initializing code cache iteration

[zhengyu123]

There are few bugs in current implementation:

1. No CodeCache_lock is held when initialize code cache iterator at safepoints. This is a problem for concurrent GC, since it can initialize a concurrent code cache iterator while holding CodeCache_lock without safepoint check.

2. Does not notify waiters upon completion of iteration at safepoints

3. Unnecessary held CodeCache_lock during concurrent code cache iteration.

Test:

• hotspot_gc_shenandoah on Linux x86_64
• hotspot_gc_shenandoah on Linux aarch64

---

Progress

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

Issue

• JDK-8276205: Shenandoah: CodeCache_lock should always be held for initializing code cache iteration

Reviewers

• Aleksey Shipilev (@shipilev - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk pull/6192/head:pull/6192
$ git checkout pull/6192

Update a local copy of the PR:
$ git checkout pull/6192
$ git pull https://git.openjdk.java.net/jdk pull/6192/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 6192

View PR using the GUI difftool:
$ git pr show -t 6192

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk/pull/6192.diff

[bridgekeeper]

👋 Welcome back zgu! 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]

@zhengyu123 The following labels will be automatically applied to this pull request:

• hotspot-gc
• shenandoah

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

[mlbridge]

Webrevs

• 00: Full (1cee3a7a)

[shipilev]

Looks okay. Do we miss _codecache_snapshot = NULL; in ~ShenandoahConcurrentRootScanner?

[openjdk]

@zhengyu123 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:

8276205: Shenandoah: CodeCache_lock should always be held for initializing code cache iteration

Reviewed-by: shade

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 111 new commits pushed to the master branch:

• c8abe35: 8276121: G1: Remove unused and uninitialized _g1h in g1SATBMarkQueueSet.hpp
• f55e68c: 8275950: Use only _thread_in_vm in ~ThreadBlockInVMPreprocess()
• 4ac8403: 8276218: JFR: Clean up jdk.jfr.dcmd
• 89ade1d: 8273416: C2: assert(false) failed: bad AD file after JDK-8252372 with UseSSE={0,1}
• dbf5100: 8276201: Shenandoah: Race results degenerated GC to enter wrong entry point
• 5bb1992: 8275868: ciReplay: Inlining fails with "unloaded signature classes" due to wrong protection domains
• 158831e: 8274320: os::fork_and_exec() should be using posix_spawn
• bf2e9ee: 8275080: G1CollectedHeap::expand() returns the wrong value
• b7104ba: 8196017: java/awt/Mouse/GetMousePositionTest/GetMousePositionWithPopup.java fails
• 6875678: 8273831: PrintServiceLookup spawns 2 threads in the current classloader, getting orphaned
• ... and 101 more: https://git.openjdk.java.net/jdk/compare/af14650127de47058b958be411503584c0ba6323...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.

[zhengyu123]

ShenandoahConcurrentRootScanner

Does it matter?

[shipilev]

ShenandoahConcurrentRootScanner

Does it matter?

You tell me, I think it would be full of dangling references after we are done with scanner? The similar reason why we do _table = NULL in a similar method in another iterator. You don't have to change it here, but maybe a simple cleanup for later.

[zhengyu123]

ShenandoahConcurrentRootScanner

Does it matter?

You tell me, I think it would be full of dangling references after we are done with scanner? The similar reason why we do _table = NULL in a similar method in another iterator. You don't have to change it here, but maybe a simple cleanup for later.

Yea, there is inconsistent among these iterators. The code is in destructors, I don't see it matters, cause accessing dangling references in a deallocated object is a fatal bug.

[zhengyu123]

/integrate

[openjdk]

Going to push as commit 99b7b95.
Since your change was applied there have been 113 commits pushed to the master branch:

• 9771544: 8260428: Drop support for pre JDK 1.4 DatagramSocketImpl implementations
• e265f83: 8276107: Preventive collections trigger before maxing out heap
• c8abe35: 8276121: G1: Remove unused and uninitialized _g1h in g1SATBMarkQueueSet.hpp
• f55e68c: 8275950: Use only _thread_in_vm in ~ThreadBlockInVMPreprocess()
• 4ac8403: 8276218: JFR: Clean up jdk.jfr.dcmd
• 89ade1d: 8273416: C2: assert(false) failed: bad AD file after JDK-8252372 with UseSSE={0,1}
• dbf5100: 8276201: Shenandoah: Race results degenerated GC to enter wrong entry point
• 5bb1992: 8275868: ciReplay: Inlining fails with "unloaded signature classes" due to wrong protection domains
• 158831e: 8274320: os::fork_and_exec() should be using posix_spawn
• bf2e9ee: 8275080: G1CollectedHeap::expand() returns the wrong value
• ... and 103 more: https://git.openjdk.java.net/jdk/compare/af14650127de47058b958be411503584c0ba6323...master

Your commit was automatically rebased without conflicts.

[openjdk]

@zhengyu123 Pushed as commit 99b7b95.

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