RFR: 8225779: Remove unused CollectedHeap::cell_header_size()

[Per Liden]

Hi,

On 6/14/19 8:45 AM, David Holmes wrote:
> Hi Per,
> 
> On 14/06/2019 4:29 pm, Per Liden wrote:
>> CollectedHeap::cell_header_size() is no longer used and can be removed. 
> 
> It seems no longer used simply because you are removing the uses. :)

Sorry, I should have been more clear. There are no longer any GC that 
overrides cell_header_size(). Shenandoah used to do that, but doesn't 
anymore.

> 
>> This patch is a hg backout of [1] where it was introduced, plus the 
>> two followup bug fixes [2] & [3].
> 
> Can you explain why you are rolling back:
> 
> "8213199: GC abstraction for Assembler::needs_explicit_null_check()"
> 
> ? Does Shenandoah no longer need it so we can just return to the simpler 
> code that existed before?

Exactly. Back when this was introduced, there were discussions about 
making "cell" and "cell header" first class concepts in the GC 
infrastructure. However, this never went further than the introduction 
of cell_header_size(). Since Shenandoah doesn't need this anymore, we 
can revert back to the old code and not have to think/talk about cells 
anymore.

> 
>> This is a clean backout, with the exception of light editing needed in 
>> two places, because:
>>
>> 1) The function Universe::narrow_oop_base() has since been renamed to 
>> CompressedOops::base().
>> 2) The include guard CPU_ARM_VM_MACROASSEMBLER_ARM_HPP has since been 
>> renamed to CPU_ARM_MACROASSEMBLER_ARM_HPP.
> 
> The backout itself looks good.

Thanks for reviewing, David!

cheers,
Per

> 
> Thanks,
> David
> ----
> 
>> Testing: Tier1-3 on all Oracle platforms
>>
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8225779
>> Webrev: http://cr.openjdk.java.net/~pliden/8225779/webrev.0
>>
>> /Per
>>
>> [1] http://hg.openjdk.java.net/jdk/jdk/rev/4ad404da0088
>> [2] http://hg.openjdk.java.net/jdk/jdk/rev/f642ede2eb07
>> [3] http://hg.openjdk.java.net/jdk/jdk/rev/18bd95c0e463