RFR: 8371716: C2 compilation fails with "Missed optimization opportunity in PhaseIterGVN"

[Marc Chevalier]

On Mon, 17 Nov 2025 09:09:33 GMT, Damon Fenacci <dfenacci at openjdk.org> wrote:

>> This bug was originally found and reported as a Valhalla problem. It quickly became apparent it has no reason to be Valhalla-specific, while I couldn't prove so. Roland managed to make a mainline reproducer. The explanation details my Valhalla investigation, but it has nothing to do with value classes anyway.
>> 
>> # Analysis
>> ## Obervationally
>> ### IGVN
>> During IGVN, in `PhiNode::Value`, a `PhiNode` has 2 inputs. Their types are:
>> 
>> in(1): java/lang/Object * (speculative=TestSpeculativeTypes$C2:NotNull:exact * (inline_depth=3))
>> in(2): null
>> 
>> We compute the join (HS' meet):
>> https://github.com/openjdk/jdk/blob/09b25cd0a24a4eaddce49917d958adc667ab5465/src/hotspot/share/opto/cfgnode.cpp#L1299-L1306
>> 
>> t=java/lang/Object * (speculative=compiler/igvn/ClashingSpeculativeTypePhiNode$C2:exact *)
>> 
>> But the current `_type` (of the `PhiNode` as a `TypeNode`) is
>> 
>> _type=java/lang/Object * (speculative=compiler/igvn/ClashingSpeculativeTypePhiNode$C1:exact *)
>> 
>> We filter `t` by `_type`
>> https://github.com/openjdk/jdk/blob/09b25cd0a24a4eaddce49917d958adc667ab5465/src/hotspot/share/opto/cfgnode.cpp#L1321
>> and we get
>> 
>> ft=java/lang/Object *
>> 
>> which is what we return. After the end of `Value`, the returned becomes the new `PhiNode`'s `_type`.
>> https://github.com/openjdk/jdk/blob/09b25cd0a24a4eaddce49917d958adc667ab5465/src/hotspot/share/opto/phaseX.cpp#L2150-L2164
>> and
>> https://github.com/openjdk/jdk/blob/09b25cd0a24a4eaddce49917d958adc667ab5465/src/hotspot/share/opto/node.cpp#L1117-L1123
>> 
>> 
>> ### Verification
>> On verification, `in(1)`, `in(2)` have the same value, so does `t`. But this time
>> 
>> _type=java/lang/Object *
>> 
>> and so after filtering `t` by (new) `_type` and we get
>> 
>> ft=java/lang/Object * (speculative=compiler/igvn/ClashingSpeculativeTypePhiNode$C2:exact *)
>> 
>> which is retuned. Verification gets angry because the new `ft` is not the same as the previous one.
>> 
>> ## But why?!
>> ### Details on type computation
>> In short, we are doing
>> 
>> t = typeof(in(1)) / typeof(in(2))
>> ft  = t /\ _type (* IGVN *)
>> ft' = t /\ ft    (* Verification *)
>> 
>> and observing that `ft != ft'`. It seems our lattice doesn't ensure `(a /\ b) /\ b = a /\ b` which is problematic for this kind of verfication that will just "try again and see if something change".
>> 
>> To me, the surprising fact was that the intersection
>> 
>> java/lang/Object * (speculative=compiler/igvn/ClashingSpeculativeTypePhiNode$C2:exact *)
>> /\
>> _type=java/lang/Objec...
>
> src/hotspot/share/opto/cfgnode.cpp line 1364:
> 
>> 1362:     const Type* first_ft = ft;
>> 1363:     ft = t->filter_speculative(first_ft);
>> 1364: #ifdef ASSERT
> 
> More of a flyby rather than a review but I was wondering if it would make sense to extract this assert block since it is the same as the one above.

We surely can. But I'd rather avoid premature clean up as it is very uncertain to me how this issue will evolve, and maybe fundamentally change.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/28331#discussion_r2533502792