8350550: Preload classes from AOT cache during VM bootstrap

[iklam]

This is still work-in-progress

---

Progress

• Change must not contain extraneous whitespace
• Change must be properly reviewed (1 review required, with at least 1 Committer)

Issue

• JDK-8350550: Preload classes from AOT cache during VM bootstrap (Enhancement - P3)

Reviewers

• John R Rose (@rose00 - Committer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/leyden.git pull/81/head:pull/81
$ git checkout pull/81

Update a local copy of the PR:
$ git checkout pull/81
$ git pull https://git.openjdk.org/leyden.git pull/81/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 81

View PR using the GUI difftool:
$ git pr show -t 81

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/leyden/pull/81.diff

[bridgekeeper]

👋 Welcome back iklam! A progress list of the required criteria for merging this PR into premain will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

[openjdk]

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

[rose00]

Very nice work. This is a Big Transaction that inhales the initial set of classes. I see it scales all the way up to the app class loader. Very cool, ship it!

Some documentation-related comments follow:

I noticed patching of the relation of class mirror to module. This is present even if the full module graph is being used. Wouldn’t it be possible to pre-configure that in the AOT cache, and only do the patching if there is a non-default module graph? Or am I misreading the code? (I’m looking at restore_module_of_preloaded_classes etc.)

The restore_unshareable_info call in SystemDictionary::preload_class is used for other miscellanous links from the class, similar to the module links I just mentioned. Same question, I guess: Why are these not pre-materialized in the AOT cache? They are “unshareable” for a reason; where is that reason documented?

And then another question: Why not handle the module linkage at the same place as the package linkage and the PD linkage? Possible answer: The module graph gets created last, so it must be back-patched. Where is that fact documented?