[sh/8u] Root processing related backports

Zhengyu Gu zgu at redhat.com
Tue Mar 17 19:00:50 UTC 2020


Adding some details:

Some of backports were not straight forward backports, especially, the 
root processor refactoring changes.

1) No WeakProcessor in 8u, ShenandoahWeakRoots just a wrapper of 
JNIHandle weak roots.

2) ShenandoahStringTableRoots is unique in 8u

3) No concurrent/single thread options from some roots. In 8u, root 
processing is always at STW.

4) Some of roots have different signatures vs. jdk/jdk, due to underline 
API differences.

5) For traversal GC, it marks all roots even when class unloading is 
enabled - followup works is needed to identify what barriers are needed.


Thanks,

-Zhengyu


On 3/17/20 12:27 PM, Zhengyu Gu wrote:
> Hi,
> 
> Please review root processing related 8u backports:
> 
> Bug list: 
> http://cr.openjdk.java.net/~zgu/shenandoah/sh-8u-root-processing/8ubackport.list 
> 
> 
> Webrev: 
> http://cr.openjdk.java.net/~zgu/shenandoah/sh-8u-root-processing/webrev.00/
> 
> 
> Test:
>     hotspot_gc_shenandoah (fastdebug and release)
>     specjvm (normal and traversal)
> 
> 
> Thanks,
> 
> -Zhengyu
> 
> 



More information about the shenandoah-dev mailing list