RFR: 8308463: Refactor regenerated class handling in lambdaFormInvokers.cpp [v2]

[Ioi Lam]

On Wed, 21 Jun 2023 00:25:46 GMT, Coleen Phillimore <coleenp at openjdk.org> wrote:

>> Ioi Lam has updated the pull request incrementally with two additional commits since the last revision:
>> 
>>  - @dholmes-ora comments
>>  - @coleenp review comments
>
> src/hotspot/share/cds/archiveBuilder.hpp line 145:
> 
>> 143:       }
>> 144:     }
>> 145:     SourceObjInfo(address src, address buf) {
> 
> This doesn't initialize all the fields, and should use initialization syntax.  Could you make one of these constructors call a common one that zeros out the other fields?

I cleaned up this constructor and added comments.

> src/hotspot/share/cds/regeneratedClasses.cpp line 75:
> 
>> 73: 
>> 74: void RegeneratedClasses::record_regenerated_objects() {
>> 75:   if (_renegerated_objs != nullptr) {
> 
> Does the DumpTimeTable_lock protect regenerated_objs also?

This function is called inside a safepoint and can't lock  DumpTimeTable_lock. I added `assert(SafepointSynchronize::is_at_safepoint()` instead.

> src/hotspot/share/cds/regeneratedClasses.hpp line 38:
> 
>> 36:   using RegeneratedObjTable = ResourceHashtable<address, address, 15889, AnyObj::C_HEAP, mtClassShared>;
>> 37:   static RegeneratedObjTable* _renegerated_objs; // InstanceKlass* and Method*
>> 38:   static GrowableArrayCHeap<OopHandle, mtClassShared>* _regenerated_mirrors;
> 
> Since these are private, and not used by the argument, do they need to be in the header file?

I moved them to the cpp file.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/14573#discussion_r1237957861
PR Review Comment: https://git.openjdk.org/jdk/pull/14573#discussion_r1237957740
PR Review Comment: https://git.openjdk.org/jdk/pull/14573#discussion_r1237957331