Re: [lworld] RFR: 8385328: [lworld] C2: assert(phi->_idx >= nodes_size()) failed: only new Phi per instance memory slice still happens after JDK-8374742 [v7]

[Roland Westrelin]

On Wed, 10 Jun 2026 21:47:26 GMT, Dean Long <dl...@openjdk.org> wrote:

>> Roland Westrelin has updated the pull request incrementally with one 
>> additional commit since the last revision:
>> 
>>   more
>
> src/hotspot/share/opto/escape.cpp line 4554:
> 
>> 4552:   if (ptn->is_JavaObject()) {
>> 4553:     Node* jobj_base = get_map(ptn->idx());
>> 4554:     if (jobj_base == nullptr || 
>> _igvn->type(jobj_base)->is_oopptr()->instance_id() != toop->instance_id()) {
> 
> I still don't like the raw instance_id() comparisons.   Could we have 
> something like _igvn->type(jobj_base)->is_oopptr()->same_instance(toop) based 
> on something like meet_instance_id() or that asserts both are 
> is_known_instance()?

I added a helper method. Does that look ok to you?

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

PR Review Comment: 
https://git.openjdk.org/valhalla/pull/2489#discussion_r3394982332