[aarch64-port-dev ] JDK 13 AArch64 issues to backport

[Aleksey Shipilev]

On 6/19/19 6:54 PM, Andrew Haley wrote:
> On 6/19/19 5:37 PM, Andrew Dinn wrote:
>> Am I supposed to backport them now or is anyone else going to step in?

>   8215879: Aarch64: ReservedStackAccess may leave stack guard in inconsistent state
> 
> No: obscure runtime option

This is not a runtime option, but the special annotation @ReservedStackAccess. But that annotation
is only available since 9 via JEP 270. So it is not applicable to 8u. Removed the affected versions.

>   8217368: AArch64: C2 recursive stack locking optimisation not triggered
> 
> No: not a correctness issue

Agreed.

>   8209413: AArch64: NPE in clhsdb jstack command
> 
> No: debugging only

I remember you were telling me that the first thing you want to always work is the debugger. It
seems to fix it, plus resolves a FIXME.

>   8216989: CardTableBarrierSetAssembler::gen_write_ref_array_post_barrier() does not check for zero length on AARCH64
> 
> No: not a correctness issue

Agreed.

>   8219993: AArch64: Compiled CI stubs are unsafely modified
> 
> No: theoretical memory-model issue, never seen in the wild

Well, kinda, do we want to bet it would not ever show up?

>   8213134: AArch64: vector shift failed with MaxVectorSize=8
> 
> No: obscure runtime option

Okay.

>   8221658: aarch64: add necessary predicate for ubfx patterns
>   8222785: aarch64: add necessary masking for immediate shift counts
> 
> No: don't get used anywhere in OpenJDK

Okay.

>   8219006: AArch64: Register corruption in slow subtype check
> 
> No: debug mode only

Same reasoning about debugging capabilities as for 8209413.


-- 
Thanks,
-Aleksey