RFR: 8146399: Refactor the BlockOffsetTable classes.

Mikael Gerdin mikael.gerdin at oracle.com
Thu Jan 7 13:25:55 UTC 2016


On 2016-01-07 14:23, Thomas Schatzl wrote:
> Hi,
>
> On Thu, 2016-01-07 at 12:27 +0100, David Lindholm wrote:
>> Hi Mikael,
>>
>> Thanks for looking at this! I have fixed your comments:
>>
>> http://cr.openjdk.java.net/~david/JDK-8146399/webrev.01/
>>
>>
>
> - the comment for the G1BlockOffsetTable constructor is out of date
> (pre-existing).
>
> - I would prefer if in g1BlockOffsetTable.inline.hpp, the definition of
> block_start_const()/block_size()/block_at_or_preceding()/forward_to_blo
> ck_containing_addr_const() the newline after the return type were
> removed.
>
> - in the same file, the arguments of block_at_or_preceding() are not
> aligned properly
>
> - in the same file, in forward_to_block_containing_addr_const(), please
> add braces for the early exit (line 140) at the if-statement.
>
> - in the same file, forward_to_block_containing() also misses braces in
> the first if-statement
>
> - g1BlockOffsetTable.cpp, additional newline in line 203
>
> - in heapRegion.inline.hpp, the change in the name of the
> G1OffsetTableContigSpace causes lots of parameter specifications to be
> not conforming to code style. Could you look over this file again?
>
> - in case the SA agent is still around, did you check the changes in
> vmstructs_g1 with it?

Just FTR, I ran David's changes and did some cursory SA testing (jmap -heap)

/Mikael

>
> Thanks a lot for the effort.
>
> Thomas
>




More information about the hotspot-gc-dev mailing list