RFR: 8240556: Abort concurrent mark after effective eager reclamation of humongous objects

Thomas Schatzl thomas.schatzl at oracle.com
Fri Sep 18 11:26:47 UTC 2020


Hi again,

On 18.09.20 12:55, Thomas Schatzl wrote:
> Hi Albert,
> 
>    thanks for your review.
> 
[...]

>  >
>  > src/hotspot/share/gc/g1/g1ConcurrentMarkThread.hpp line 112:
>  >
>  >> 110:   bool idle() const;
>  >> 111:   bool started() const;
>  >> 112:   bool in_progress() const;
>  >
>  > What's the motivation of moving them out of the header? Since they are
>  > quite small, placing them in the header makes
>  > inlining possible and desirable.
> 
> This way the implementation does not distract from reading the .hpp file 
> as they've become larger.
> 
> There is no gain in inlining them imho as the callers are not 
> performance sensitive (as in: this is called once now and then), so I 
> did not add a .inline.hpp file.

Actually there is a .inline.hpp file already there. I'll add these 
methods there although I think it does not matter performance-wise; 
these are called once per marking cycle, and most of the time from the 
.cpp file.

Thanks,
   Thomas


More information about the hotspot-jfr-dev mailing list