RFR: 8356159: RISC-V: Add Zabha [v2]

[Robbin Ehn]

On Fri, 16 May 2025 07:03:20 GMT, Fei Yang <fyang at openjdk.org> wrote:

>> Robbin Ehn has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains four additional commits since the last revision:
>> 
>>  - Revert back to default relaxed
>>  - Merge branch 'master' into 8356159
>>  - Fixed ws
>>  - Initial draft
>
> src/hotspot/cpu/riscv/macroAssembler_riscv.cpp line 3842:
> 
>> 3840:                                                  Register shift, Register mask, Register aligned_addr) {
>> 3841:   assert(size == int8 || size == int16, "unsupported operand size");
>> 3842:   assert(!(UseZacas && UseZabha), "Use amocas");
> 
> Seems more obvious to move this two assertions to the entry of its callers?
> `MacroAssembler::cmpxchg_narrow_value` and `MacroAssembler::weak_cmpxchg_narrow_value`.

The sizing checking one we really need as otherwise this method generates broken code.
Fixed.

> src/hotspot/cpu/riscv/riscv.ad line 5820:
> 
>> 5818:   ins_encode %{
>> 5819:     __ cmpxchg(as_Register($mem$$base), $oldval$$Register, $newval$$Register, Assembler::int8,
>> 5820:               /*acquire*/ Assembler::aq, /*release*/ Assembler::rl, $res$$Register);
> 
> Indentation.

Looks like the other one below:

  ins_encode %{
    __ cmpxchg_narrow_value(as_Register($mem$$base), $oldval$$Register, $newval$$Register, Assembler::int16,
                            /*acquire*/ Assembler::aq, /*release*/ Assembler::rl, $res$$Register,
                            /*result_as_bool*/ false, $tmp1$$Register, $tmp2$$Register, $tmp3$$Register);
  %}


No ?

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

PR Review Comment: https://git.openjdk.org/jdk/pull/25252#discussion_r2097797113
PR Review Comment: https://git.openjdk.org/jdk/pull/25252#discussion_r2097793856