RFR: 8355915: [leyden] Crash in MDO clearing the unloaded array type

[Aleksey Shipilev]

On Tue, 29 Apr 2025 14:57:23 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

> Caught this when doing benchmarks with Spring Boot. See the bug for reproducer.
> 
> The lifecycle of array types of `T` is bound to the lifecycle of `T` themselves. So the fix does similar thing to `is_excluded` and other code: when we encounter the `T[]`, we ask if its bottom component classes is not yet initialized. This fits nicely with our current uses, which skip such classes, and ever touch their (potentially nullptr) CLDs.
> 
> Additional testing:
>  - [x] Ad-hoc testing with most recent spring-boot-petclinic
>  - [x] Linux x86_64 server fastdebug, `runtime/cds`

Testing shows the fix is incomplete. 

Looks like when `T[]` is unloaded, we cannot trust its `bottom_klass` is not garbage. I also notice that in several other places, e.g. in `ciReceiverTypeData::translate_receiver_data_from`, we do check if CLD is null. So the more robust fix is to do CLD checks in MDO clearing as well. `is_excluded` is now also protected by these CLD checks.

See new commit. This one seems to pass the tests locally.

-------------

PR Comment: https://git.openjdk.org/leyden/pull/64#issuecomment-2841330187