RFR: 8325467: Support methods with many arguments in C2 [v12]

[Daniel Lundén]

On Tue, 1 Apr 2025 16:16:36 GMT, Roberto Castañeda Lozano <rcastanedalo at openjdk.org> wrote:

>> Daniel Lundén has updated the pull request incrementally with two additional commits since the last revision:
>> 
>>  - Formatting updates
>>  - Add register mask fuzzer test
>
> src/hotspot/share/opto/chaitin.cpp line 1655:
> 
>> 1653:       // Bump register mask up to next stack chunk
>> 1654:       bool success = lrg->rollover();
>> 1655:       if (!success) {
> 
> Was this scenario (running out of stack slots representable in `OptoRegPairs`) possible before, or was it prevented by some check removed in the changeset? Did you come across it in some compilation or is it more of a "theoretical" guard?

Yes, it is a theoretical guard (also see the discussions earlier in this PR) and could also happen before this changeset if we roll over too much in `Select`. I experimented a bit with this earlier on and was not able to construct an example where we end up in this situation.

> src/hotspot/share/opto/regmask.hpp line 282:
> 
>> 280:       _grow(src._rm_size, false);
>> 281:       memcpy(_RM_UP_EXT, src._RM_UP_EXT,
>> 282:              sizeof(uintptr_t) * (src._rm_size - _RM_SIZE));
> 
> This code is not very well covered by current tests, please consider adding some tests to `test_regmask.cpp` to exercise it.

Now added!

> src/hotspot/share/opto/regmask.hpp line 293:
> 
>> 291:         _hwm = _rm_max();
>> 292:       }
>> 293:       _set_range(src._rm_size, value, _rm_size - src._rm_size);
> 
> This code is not very well covered by current tests, please consider adding some tests to `test_regmask.cpp` to exercise it.

Now added!

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

PR Review Comment: https://git.openjdk.org/jdk/pull/20404#discussion_r2026941355
PR Review Comment: https://git.openjdk.org/jdk/pull/20404#discussion_r2026941803
PR Review Comment: https://git.openjdk.org/jdk/pull/20404#discussion_r2026941926