Review Request: 6896043: Zero fixes

Vladimir Kozlov Vladimir.Kozlov at Sun.COM
Mon Nov 23 10:06:26 PST 2009


Gary Benson wrote:
> Hi Vladimir,
> 
> Vladimir Kozlov wrote:
>>>   hotspot/src/cpu/zero/vm/entry_zero.hpp:
>> To be clear. You did it to inline methods in debug VM version,
>> right?  They will be optimized for optimized builds anyway.
> 
> No, I did it to inline them for the optimized build.  I'd assumed
> that methods defined in header files like that would be inlined,
> but they don't seem to be.  Maybe this is a peculiarity of the
> IcedTea build, which is where I noticed the problem.  If it is,
> I can remove it from this webrev.

As Andrew said, functions defined inside class definition should
be inlined by default. If that not happened then adding "inline"
will not help. I was mostly surprise that you add "inline" for
the method which just return a field value.

But if it really helped (you verified the code) then leave it.

>>>   hotspot/src/share/vm/runtime/os.hpp:
>> Can you explain why your changes is not the same as the comment
>> says?: ((_mem_serialize_page ^ addr) & -pagesize) == 0
> 
> Basically because I didn't know if I'd need to make changes anywhere
> else, and I didn't want to break the other platforms.  Should I change
> it to what the comment says?

I am concern about correctness of your code - page sizes could be different.
I would prefer if your code will be similar to one in is_poll_address().

Vladimir

> 
> Cheers,
> Gary
> 


More information about the hotspot-runtime-dev mailing list