[PATCH] openjdk/zero: only use FPU opcodes if FPU is available
tom.rodriguez at oracle.com
Wed Jun 23 10:24:58 PDT 2010
On Jun 23, 2010, at 2:22 AM, Gary Benson wrote:
> Sebastian Andrzej Siewior wrote:
>> * Gary Benson | 2010-06-23 09:29:21 [+0100]:
>>> Basically, if you just want to run stuff then it's ok to bypass
>>> that piece of assembler, but if you want to certify the builds
>>> for whatever reason then you will need to figure out some other
>>> way to do a 64-bit atomic copy.
>> Is it garanteed that the adress I'm reading from and writing to is
>> properly aligned on a 64bit address boundary?
> Not sure. I guess that's a question for one of the HotSpot guys:
> are long and double fields 64-bit aligned on 32-bit machines? Or,
> if not, is it possible to make them be?
Yes they are 64 bit aligned. They have to be for volatile longs and doubles to work efficiently.
>>>>> Have you signed the SCA? If so I can (or you can) make a
>>>>> webrev for this and submit it for inclusion.
>>>> I haven't.
>>> Are you able to sign it?
>> I clicked that link but I ended somewhere at oracle.com with 404.
> So it does. Maybe someone from Oracle could point you to the correct
More information about the hotspot-compiler-dev