RFR(S) 8238782: Cleanup Deoptimization::deoptimize(): remove unused RegisterMap argument and don't update RegisterMap in callers if UseBiasedLocking is enabled

Reingruber, Richard richard.reingruber at sap.com
Wed Feb 12 08:44:29 UTC 2020


Hi David,

thanks for reviewing!

Richard.

-----Original Message-----
From: David Holmes <david.holmes at oracle.com> 
Sent: Mittwoch, 12. Februar 2020 04:40
To: Reingruber, Richard <richard.reingruber at sap.com>; hotspot-dev at openjdk.java.net; hotspot-runtime-dev at openjdk.java.net
Subject: Re: RFR(S) 8238782: Cleanup Deoptimization::deoptimize(): remove unused RegisterMap argument and don't update RegisterMap in callers if UseBiasedLocking is enabled

Hi Richard,

This cleanup looks good to me!

Thanks,
David

On 11/02/2020 10:52 pm, Reingruber, Richard wrote:
> Hi,
> 
> I would like to do a small clean-up removing the unused parameter "map" from
> Deoptimization::deoptimize(JavaThread* thread, frame fr, RegisterMap *map, DeoptReason reason).
> 
> Additionally the callers are changed to use a non-updating RegisterMap to iterate the stack
> frames. The updating RegisterMap was needed only for revoking biases in
> Deoptimization::deoptimize(), and since JDK-8226705 this isn't done in deoptimize() anymore[1][2][3].
> 
> So please review
> 
> Webrev: http://cr.openjdk.java.net/~rrich/webrevs/8238782/webrev.0/
> Bug:    https://bugs.openjdk.java.net/browse/JDK-8238782
> 
> Testing: JCK and JTREG tests, also in Xcomp mode with fastdebug and release builds on all platforms.
> 
> Thanks, Richard.
> 
> See also:
> 
> [1] JDK-8226705 removes call to revoke_biases_of_monitors() from Deoptimization::deoptimize()
>      http://hg.openjdk.java.net/jdk/jdk/rev/408c445d04e8#l21.320
> 
> [2] JDK-8226705 adds call to revoke_from_deopt_handler() to Deoptimization::fetch_unroll_info_helper() [3]
>      http://hg.openjdk.java.net/jdk/jdk/rev/408c445d04e8#l21.199
> 
> [3] JDK-8226705: Deoptimization::revoke_from_deopt_handler() revokes biases just before replacing a
>      compiled frame with equivalent interpreter frames.
>      http://hg.openjdk.java.net/jdk/jdk/rev/408c445d04e8#l21.290
> 


More information about the hotspot-runtime-dev mailing list