C2: Understanding LoadNode::find_previous_arraycopy()

[Reingruber, Richard]

Hi Vladimir,

Yes, I noticed this too.

Thanks, Richard.

-----Original Message-----
From: Vladimir Kozlov <vladimir.kozlov at oracle.com> 
Sent: Mittwoch, 17. März 2021 17:11
To: Reingruber, Richard <richard.reingruber at sap.com>; Roland Westrelin <rwestrel at redhat.com>; hotspot-compiler-dev at openjdk.java.net
Subject: Re: C2: Understanding LoadNode::find_previous_arraycopy()

Original 8076188 changes returned `ac`:

https://github.com/openjdk/jdk/commit/a9cdbd04076149927fc7c13704eb01ea32e2ca6f#diff-819eefad0e0c83dec6968dc3f74924e3cd1763f3a43edf305f37ff591ad9a2feR562

but then 8130847 changes accidently removed it when code was refactored:

https://github.com/openjdk/jdk/commit/0baf2f7e8ab034548c927857c3182cc949bb225b#diff-819eefad0e0c83dec6968dc3f74924e3cd1763f3a43edf305f37ff591ad9a2feL567

Thanks,
Vladimir K

On 3/17/21 8:27 AM, Reingruber, Richard wrote:
> Hi Roland,
> 
>>> Should we return ac when executing L556?
> 
>> Yes, I think that was the intent.
> 
> Thanks, I'll do some testing.
> 
> Richard.
> 
> -----Original Message-----
> From: Roland Westrelin <rwestrel at redhat.com>
> Sent: Mittwoch, 17. März 2021 13:51
> To: Reingruber, Richard <richard.reingruber at sap.com>; hotspot-compiler-dev at openjdk.java.net
> Subject: Re: C2: Understanding LoadNode::find_previous_arraycopy()
> 
> 
> Hi Richard,
> 
>> Should we return ac when executing L556?
> 
> Yes, I think that was the intent.
> 
> Roland.
>