8308745: ObjArrayKlass::allocate_objArray_klass may call into java while holding a lock

[coleenp]

This change uses a claim token to allocate multi dimensional arrays rather than holding MultiArray_lock around metaspace allocation. We can't hold a mutex around metaspace allocation because it can create an OOM object and it can also call into JVMTI for a resource exhausted event. Also, we were creating mirrors and more metadata arrays while holding this lock. See the bug for more details and other ideas considered and rejected.

Tested with tier1-7.

---

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-8308745: ObjArrayKlass::allocate_objArray_klass may call into java while holding a lock (Bug - P4)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/17660/head:pull/17660
$ git checkout pull/17660

Update a local copy of the PR:
$ git checkout pull/17660
$ git pull https://git.openjdk.org/jdk.git pull/17660/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 17660

View PR using the GUI difftool:
$ git pr show -t 17660

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/17660.diff

Webrev

Link to Webrev Comment

[bridgekeeper]

👋 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]

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

• hotspot
• serviceability

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 (403dd830)

[dholmes-ora]

Can't you just use a Monitor to implement the claim token, rather than this lock-free approach? (Similar to how class initialization is handled.)

Sorry lost the forest in the trees.

[coleenp]

I'm going to do a bit of rework as discussed with @dholmes-ora .