Backport: 8216549: Mismatched unsafe access to non escaping object fails

[Hohensee, Paul]

Stuff gets lost in mailboxes all the time. :)

Paul

On 10/10/19, 8:55 AM, "jdk8u-dev on behalf of Hohensee, Paul" <jdk8u-dev-bounces at openjdk.java.net on behalf of hohensee at amazon.com> wrote:

    We're starting to have a general problem, which is that as the tip code base diverges from the update code base, we're getting more backports that depend on other backports. At some point not doing those other backports is going to cause reliability problems.
    
    In this particular case it's pretty obvious we're ok, because as you say, the extra 2 lines are just completing the 8162101 backport. Having to leave out code in a backport such as 8162101 is a sign of a problem: it should have been addressed then.
    
    Anyway, this patch is good to go afaic.
    
    Thanks,
    
    Paul
    
    On 10/10/19, 3:19 AM, "Roman Kennke" <rkennke at redhat.com> wrote:
    
        Hi Paul,
        
        > Something of a nit: most of 8162101 is indeed in 8u :), but the change at line 1.42/1.43
        > 
        > -  } else if (alias_type->adr_type() == TypeOopPtr::BOTTOM) {
        > +  } else if (alias_type->adr_type()->isa_oopptr()) {
        > 
        > is missing because the test wasn't already there.
        
        Right.
        
        I tried backporting JDK-8160360, but the problem there is similar: in
        order to be a clean-ish backport, it requires a couple of other changes
        first. That's a rat's tail of (unrelated/unimportant) stuff to backport
        just for those two lines. Instead, I'd rather opt to include the above 2
        lines (which are essentially completing the JDK-8162101 backport) instead.
        
        Roman
        
        
        
        
        > Thanks,
        > 
        > Paul
        > 
        > On 10/9/19, 3:07 AM, "Roman Kennke" <rkennke at redhat.com> wrote:
        > 
        >     Hi Paul,
        >     
        >     Thanks for checking it!
        >     
        >     > The code in library_call.cpp comes seems to come from two patches
        >     > 
        >     > https://bugs.openjdk.java.net/browse/JDK-8160360
        >     > http://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/4be0cada20ad
        >     
        >     I will look into backporting this change.
        >     
        >     > https://bugs.openjdk.java.net/browse/JDK-8162101
        >     > http://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/10dad1d40843
        >     
        >     This is already in jdk8u:
        >     http://hg.openjdk.java.net/jdk8u/jdk8u-dev/hotspot/rev/7ca49bca3c2a
        >     
        >     Roman
        >     
        >     
        >     
        >     > I'd backport those first.
        >     > 
        >     > Paul
        >     > 
        >     > On 10/7/19, 11:31 AM, "jdk8u-dev on behalf of Roman Kennke" <jdk8u-dev-bounces at openjdk.java.net on behalf of rkennke at redhat.com> wrote:
        >     > 
        >     >     Ping! This is still awaiting review.
        >     >     
        >     >     Thanks,
        >     >     Roman
        >     >     
        >     >     
        >     >     > This backports:
        >     >     > https://bugs.openjdk.java.net/browse/JDK-8216549
        >     >     > 
        >     >     > Original change:
        >     >     > hg.openjdk.java.net/jdk/jdk12/rev/f0490430ef7a
        >     >     > 
        >     >     > JDK11u backport:
        >     >     > http://hg.openjdk.java.net/jdk-updates/jdk11u-dev/rev/d5fb6ad4203b
        >     >     > 
        >     >     > In order to make it work on 8u, I needed the extra change in
        >     >     > library_call.cpp. Thanks Roland who helped with this.
        >     >     > 
        >     >     > Webrev:
        >     >     > http://cr.openjdk.java.net/~rkennke/JDK-8216549-8u/webrev.00/
        >     >     > 
        >     >     > Testing:
        >     >     > The new testcase fails without the fix, and passes with the fix. tier1 &
        >     >     > tier2 show no regressions.
        >     >     > 
        >     >     > Can I please get a review?
        >     >     > 
        >     >     > Thanks,
        >     >     > Roman
        >     >     > 
        >     >     
        >     >     
        >     > 
        >     
        >     
        >