RFR: JDK-8222259: Shenandoah: Pre-evacuate string-dedup roots in Traversal GC

[Roman Kennke]

> On 4/10/19 1:28 PM, Roman Kennke wrote:
> > Fix:
> > http://cr.openjdk.java.net/~rkennke/JDK-8222259/webrev.00/
> 
> It looks okay, but I would have expected the call to
> ShenandoahRootProcessor, rather than naked root
> walk. Isn't there a path into SRP that does it for traversal?

Currently not, and the whole idea of pre-evacing anything from
traversal is a bit of a cludge currently. For example, we need to pre-
evac everything from code-roots currently, even if we stricly only want
strong stuff marked. The solution would be to pre-evac strong code
roots via stacks, and then let nmethod barrier pre-evac everything ever
entered during the cycle. All of this requires deeper looking into it,
but probably later when we're seriously into stabilizing traversal GC.

Roman

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20190410/3bff7985/signature.asc>