[aarch64-port-dev ] [11] RFR(S) 8202710: sporadic jtreg test fail on aarch64

[Boris Ulasevich]

Dmitry, thanks!
[adding hotspot-compiler-dev]

Hello all,

Can you please review my test. Basically my change is about aarch64 
issue, but I added jtreg test which goes to shared code, so I need 
additional approval.

Bug: https://bugs.openjdk.java.net/browse/JDK-8202710
Fix: http://cr.openjdk.java.net/~bulasevich/8202710/webrev.01

Thank you,
Boris

On 10.05.2018 15:19, Dmitry Samersoff wrote:
> Boris,
> 
> The fix looks OK to me.
> 
> -Dmitry
> 
> On 09.05.2018 09:53, Boris Ulasevich wrote:
>> Hi,
>>
>> Please review the following change.
>>
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8202710
>> Fix: http://cr.openjdk.java.net/~bulasevich/8202710/webrev.01
>>
>> Aarch64 deopt_blob stores float values to stack as is, after that
>> deoptimizer treats values as float-converted-to-double - this makes
>> values being wrongly restored after safe point handler.
>>
>> The fix is quite evident, just correcting float_in_double setting in
>> AArch64 Description File. New jtreg tests provided to reproduce the
>> issue stably. By the way the error in compile2runtime handler blob fixed
>> to avoid possible save/restore registers number mismatch (I have no test
>> to reproduce the problem).
>>
>> thanks,
>> Boris