RFR 8229919: Support JNI Critical functions in object pinning API on x86_32 platforms

[Zhengyu Gu]

Thanks, Andrew.

I will add comments as you suggested before push.

-Zhengyu

On 10/15/19 10:52 AM, Andrew Dinn wrote:
> Hi Zhengyu,
> 
> On 04/10/2019 20:22, Zhengyu Gu wrote:
>> Ping! May I get a second review?
> 
> Yes, the patch looks good.
> 
> Only one tentative suggestion. It might help those coming to this from
> the x86_64 code to add a comment before the calls to gen_pin_object and
> gen_unpin_object to mention that those calls handle saving and restoring
> of any registers clobbered by the call i.e.
> 
> 1986           if (Universe::heap()->supports_object_pinning()) {
> +                 // gen_pin_object handles save and restore
> +                 // of any clobbered registers
> 1987             gen_pin_object(masm, thread, in_arg);
> 
> and
> 
> 
> 2209       }
> +                 // gen_pin_object handles save and restore
> +                 // of any other clobbered registers
> 2210       gen_unpin_object(masm, thread, in_regs[i]);
> 
> (note 'other' in the 2nd comment means other than the result register)
> 
> You decide. I don't need another webrev.
> 
> regards,
> 
> 
> Andrew Dinn
> -----------
> 
>> On 9/18/19 10:46 AM, Roman Kennke wrote:
>>> Hi Zhengyu,
>>>
>>> It looks good to me!
>>>
>>> Thanks,
>>> Roman
>>>
>>>
>>>> Please review this patch that supports JNI critical functions in
>>>> object pinning capable GCs on x86_32 platforms.
>>>>
>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8229919
>>>> Webrev: http://cr.openjdk.java.net/~zgu/JDK-8229919/webrev.00/
>>>>
>>>> Test:
>>>>     hotspot_gc_shenandoah (fastdebug and release) with 32-bit VM on
>>>> Linux x86_64.
>>>>     hotspot_gc, hotspot_runtime and hotspot_compiler
>>>>     Submit tests in progress.
>>>>
>>>> Thanks,
>>>>
>>>> -Zhengyu
>>