[lworld] RFR: 8372824: [lworld] C2 hits "Unexpected argument type" assertion with --enable-preview [v2]

[Quan Anh Mai]

On Tue, 20 Jan 2026 16:23:08 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

>> Quan Anh Mai has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Unproblemlist
>
> src/hotspot/share/opto/graphKit.cpp line 1960:
> 
>> 1958:           // scalarized receiver instead.
>> 1959:           assert(arg_idx == 0 && !call->method()->is_static(), "must be the receiver");
>> 1960:           arg = InlineTypeNode::make_from_oop(this, arg, t->inline_klass());
> 
> So this is post-parse call-devirtualization, i.e. [JDK-8257211](https://bugs.openjdk.org/browse/JDK-8257211), right? Could we assert `call->generator()->is_virtual_late_inline()` here?

Unfortunately, the `CallNode` here does not have a generator, because it is created just above. Checking `is_late_inline` does not work either, because we are too deep handling now. `LateInlineVirtualCallGenerator::do_late_inline_check` calls into `Compile::call_generator` to find the replacement. This method, after deciding that it does not want to inline the callee, will create a  `PredictedCallGenerator` with the fast path being a `DirectCallGenerator`. And this `DirectCallGenerator` is the one calling the method we are in, and this subclass of `CallGenerator` does not return `true` for `is_late_inline()`.

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

PR Review Comment: https://git.openjdk.org/valhalla/pull/1935#discussion_r2709698942