RFR: 8302215: G1: Last-ditch Full GC should do serial compaction for tail regions in per thread compaction points. [v5]

[Albert Mingkun Yang]

On Fri, 17 Feb 2023 09:41:57 GMT, Ivan Walulya <iwalulya at openjdk.org> wrote:

>> Hi all, 
>> 
>> Please review this change to call `G1FullCollector::phase2c_prepare_serial_compaction` when `do_maximal_compaction`. For maximum compaction, we re-prepare all objects in the "tail" regions of the per thread compaction points.
>> 
>> Testing: Tier 1-3.
>> 
>> Thanks
>
> Ivan Walulya has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Thomas suggestions

src/hotspot/share/gc/g1/g1FullCollector.cpp line 396:

> 394:         // since it is already prepared for compaction.
> 395:         serial_cp->initialize(current);
> 396:         dense_prefix_top = compaction_top(current);

I think `serial_cp->initialize(current);` and `dense_prefix_top` can be extracted out the loop, because they correspond to `lowest_current`, which is "statically" known before entering the loop.

src/hotspot/share/gc/g1/g1FullGCPrepareTask.hpp line 124:

> 122:   HeapRegion* _current;
> 123:   HeapWord* _dense_prefix_top;
> 124:   uint _bottom_index;

Seems unused.

src/hotspot/share/gc/g1/g1FullGCPrepareTask.hpp line 129:

> 127:   G1SerialRePrepareClosure(G1FullGCCompactionPoint* hrcp, HeapRegion* hr, HeapWord* dense_prefix_top) :
> 128:     _cp(hrcp),
> 129:     _current(hr),

Can't find where it's used.

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

PR: https://git.openjdk.org/jdk/pull/12529