RFR: 8352681: C2 compilation hits asserts "must set the initial type just once"

[Christian Hagedorn]

On Sun, 6 Apr 2025 04:04:56 GMT, Cesar Soares Lucas <cslucas at openjdk.org> wrote:

> I didn't opt to make _delay_transform false during the RAM transformations because that seemed to be a risky move, nevertheless I'll create an RFE and keep investigating that option.
> 
> Also, while working on a test case I found that C2 doesn't remove (at least not before 
EA steps) the if (param != param) {...} and I'm going to investigate that as a separate RFE.

Sounds good!

The fix looks good to me.

src/hotspot/share/opto/memnode.cpp line 293:

> 291:         toop->offset() == Type::OffsetBot)) {
> 292:     // IGVN _delay_transform may be set to true and if that is the case and mmem
> 293:     // is already a registered then the validation inside transform will

Suggestion:

    // is already a registered node then the validation inside transform will

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

Marked as reviewed by chagedorn (Reviewer).

PR Review: https://git.openjdk.org/jdk/pull/24471#pullrequestreview-2745721433
PR Review Comment: https://git.openjdk.org/jdk/pull/24471#discussion_r2030581836