RFR (T) 8243393: Improve ReservedSpace constructor resolution

David Holmes david.holmes at oracle.com
Wed Apr 29 03:53:19 UTC 2020

Hi Coleen,

On 28/04/2020 2:12 am, coleen.phillimore at oracle.com wrote:
> Summary: Remove possibly ambiguous constructor and use directly in 
> ReservedCodeHeap
> Tested with tier1-3.
> open webrev at http://cr.openjdk.java.net/~coleenp/2020/8243393.01/webrev
> bug link https://bugs.openjdk.java.net/browse/JDK-8243393

Sorry but I don't understand the mapping to the constructors. In this 
original call from cardTable.cpp:

ReservedSpace heap_rs(_byte_map_size, rs_align, false);

this calls the 4-arg constructor as-if:

ReservedSpace heap_rs(_byte_map_size, rs_align, false, NULL);

and rs_align maps to the alignment parameter.

But in the new two-arg constructor call:

ReservedSpace heap_rs(_byte_map_size, rs_align)

the rs_align is no longer treated as alignment but rather 
preferred_page_size! It isn't at all obvious to me that this is an 
equivalent construction.

Both constructors need to be adequately documented - at present only the 
two-arg constructor is.


> Thanks,
> Coleen

More information about the hotspot-runtime-dev mailing list