RFR: 8341976: C2: use_mem_state != load->find_exact_control(load->in(0)) assert failure [v2]

[Roland Westrelin]

On Mon, 24 Mar 2025 12:02:31 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:

>>> Drive by comments: Is `-XX:-UseOnStackReplacement` required to reproduce the issue?
>> 
>> No, it's not. I trimmed the list of options for the tests.
>> 
>>> There was also a crash when running with `-XX:+TraceLoopOpts`. Can you also add a run with that flag to verify that this patch also fixes that?
>> 
>> Added. The `TraceLoopOpts` crash reproduces: the code hits a malformed counted loop. I tweaked the printing code.
>
>> Added. The TraceLoopOpts crash reproduces: the code hits a malformed counted loop. I tweaked the printing code.
> 
> Is the malformed counted loop expected or a different issue to look into?

It doesn't look an actual issue to me. `PhiNode::Value` manages to narrow the trip `phi`'s type of the pre loop enough that it's a constant. So the loop no longer has the expected counted loop shape but the loop exit condition that should constant fold doesn't because it's guarded by an `Opaque1` node.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/23465#discussion_r2018941323