RFR: 8325885: GenShen: harmonize with JDK-8325671

[Y. Srinivas Ramakrishna]

On Mon, 4 Mar 2024 21:43:50 GMT, Kelvin Nilsen <kdnilsen at openjdk.org> wrote:

>> /issueJDK-8325885
>
> src/hotspot/share/gc/shenandoah/shenandoahConcurrentMark.cpp line 244:
> 
>> 242:   ShenandoahFlushSATBHandshakeClosure flush_satb(qset);
>> 243:   for (uint flushes = 0; flushes < ShenandoahMaxSATBBufferFlushes; flushes++) {
>> 244:     // TODO: ysr : hoist this switch out of the loop & refactor ?? How may
> 
> The default value of ShenandoahMaxSATBBufferFlushes is 5.  I think the number of iterations is typically low.  Given low number of typical iterations, may not be worth it to restructure the loop control.

Good point, I'll clear this comment which was inadvertently left behind here.

To your point, the cost of the switch is far far lower than the work done inside each case statement, so any fears of cost when I initially wrote that comment are overblown.

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

PR Review Comment: https://git.openjdk.org/shenandoah/pull/402#discussion_r1513247442