RFR: sh/jdk11 assert failure because IfNode input is a Phi

[Aleksey Shipilev]

On 8/28/19 4:45 PM, Roland Westrelin wrote:
>> I was trying to write a test case for that bug so I could fix it in
>> jdk/jdk and then I realized that the fix above was addressing the
>> consequence of logic that was missing and should have been
>> backported. So I'd like to back it out and backport what's missing
>> instead:
>>
>> http://cr.openjdk.java.net/~roland/shenandoah/LRBafterCallUseBool/webrev.01/
> 
> Anyone to approve this?

This seems to be partially backporting this change, right?
  8213381: Hook to allow GC to inject Node::Ideal() calls

We would need to pick that up in 11.0.6 to minimize upstream differences. The current patch seems
okay for sh/jdk11, as it is pretty straight-forward. If we want to be extra-clean, I'd wait for
8213381 backport to arrive via 11u.

-- 
Thanks,
-Aleksey