RFR: 8339870: Remove yet more stale TODOs

[Y. Srinivas Ramakrishna]

On Tue, 10 Sep 2024 19:16:39 GMT, Y. Srinivas Ramakrishna <ysr at openjdk.org> wrote:

>> Final round of TODO removal.
>> * Move asterisk for pointer next to type, not variable
>> * Remembered set verifier uses a small 'scanner' abstraction to simplify reading of read or write card table
>> 
>> ## Testing
>> 
>> GHA, Internal pipelines
>
> src/hotspot/share/gc/shenandoah/shenandoahConcurrentGC.cpp line 687:
> 
>> 685:     _generation->prepare_regions_and_collection_set(true /*concurrent*/);
>> 686: 
>> 687:     // Upon return from prepare_regions_and_collection_set(), certain parameters have been established to govern the
> 
> For the block comment at lines 687-702 (right pane), would it make sense to move it to the spec of `prepare_...` and remove this block comment from here?

In fact, these are already present in the documentation for the method, so we should just remove this and declutter the code here.

> src/hotspot/share/gc/shenandoah/shenandoahScanRemembered.cpp line 91:
> 
>> 89: }
>> 90: 
>> 91: bool ShenandoahDirectCardMarkRememberedSet::is_write_card_dirty(HeapWord* p) const {
> 
> Interesting. Is this new? Is it used somewhere or introduced for the purpose of a more consistent/uniform API?
> 
> (Not requiring any change; just asking. I am all for uniform/consistent APIs.)

Never mind. Found the (new) place where it's now used.

> src/hotspot/share/gc/shenandoah/shenandoahScanRemembered.cpp line 324:
> 
>> 322: }
>> 323: 
>> 324: bool ShenandoahScanRemembered::is_write_card_dirty(HeapWord* p) {
> 
> Same question/remark as I asked/made for the ShenDirectCardMarkRS api above at line 91.

Never mind; found where it's used now.

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

PR Review Comment: https://git.openjdk.org/shenandoah/pull/496#discussion_r1752638441
PR Review Comment: https://git.openjdk.org/shenandoah/pull/496#discussion_r1752690061
PR Review Comment: https://git.openjdk.org/shenandoah/pull/496#discussion_r1752690555