[11u] RFC 8231405: [Shenandoah] guarantee(d != NULL) failed: Null dominator info

Andrew John Hughes gnu.andrew at redhat.com
Mon Sep 30 19:22:28 UTC 2019



On 30/09/2019 13:14, Aleksey Shipilev wrote:
> Hi,
> 
> This is the question for Andrew Hughes, as we have reached the CPU work already. We have this
> critical bugfix in Shenandoah/C2 that we would like to have in 11.0.5:
>   https://bugs.openjdk.java.net/browse/JDK-8231405
>   https://hg.openjdk.java.net/jdk/jdk/rev/222a91b9438a
> 
> The patch applies cleanly, but we are usually prepend the synopsis with "[backport]" to disambiguate
> from the jdk-updates/jdk11u backports:
>   "[backport] 8231405: [Shenandoah] guarantee(d != NULL) failed: Null dominator info"
> 
> I ran hotspot_gc_shenandoah, tier1, tier2 (with Shenandoah), jdk_security (with Shenandoah).
> 
> Can we have it? If yes, there are several ways to do this:
> 
>  a) Push the changeset to CPU forests directly, and then wait for the entire bundle to appear in
> sh/jdk11 with the CPU patchset. Pros: it would get into CPU without additional fuzz. Cons: it would
> not get tested with our public 11u nightlies, until the CPU patchset arrives.
> 
>  b) Push the changeset to sh/jdk11, pull the sh/jdk to CPU forests. Pros: it would be exposed to
> users before the CPU release through our 11u nightlies. Cons: the later CPU merge would be against
> that public changeset, not against shenandoah-jdk-11.0.5+9.
> 
> Thoughts?
> 

The second option, b, makes sense to me. I don't really follow the con
for that; there are plenty of instances where Shenandoah changes have
gone in between merges. I think a bit of noise on a webrev is
outweighted by the testing and visibility it would get in the public repos.

As to the viability of the fix itself, it's Shenandoah-only code that it
affects, so I'll leave that call up to you.

Thanks,
-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222
https://keybase.io/gnu_andrew



More information about the shenandoah-dev mailing list