Peek through barriers in Node::eqv_uncast()

[Aleksey Shipilev]

On 10/17/2018 04:29 PM, Roman Kennke wrote:
>>> Yes, indeed. But let's spin it a bit in Shenandoah first, shall we?
>>> In-fact, I want to run the lucene suite a few more times before even
>>> pushing it, each run takes 1-2hours so maybe until EOD? This frickin bug
>>> tends to be very spurious.
>>
>> Regardless if it fixes the current bug, it look as general thing to deal with for all GCs. So, why
>> wait for those results? Maybe pass it via jdk-submit and RFR upstream directly?
>>
>>> http://cr.openjdk.java.net/~rkennke/c2-eqv/webrev.01/
>>
>> Looks good to me. The comment needs to mention GC barriers too?
>>
>>  463   // Return whether two Nodes are equivalent, after stripping casting.
>>  464   bool eqv_uncast(const Node* n) const;
>>
>> Renaming the method would probably be quite intrusive...
> 
> I added a comment and will push it as cherry-pick for:
> https://bugs.openjdk.java.net/browse/JDK-8212603

Fine by me.

-Aleksey