RFR: 8317720: RISC-V: Implement Adler32 intrinsic [v15]

[Fei Yang]

On Wed, 10 Jul 2024 14:11:33 GMT, ArsenyBochkarev <duke at openjdk.org> wrote:

> Sorry for such a late reply. It is a good catch, I see it, thanks! However, it is not correct for all cases: consider having -63 after `L_by16_loop_unroll`

Hmm ... I don't understand how this could happen. I see `len >= 64` is ensured when we first enter `L_by16_loop_unroll`. And you do `len -= 64` and compare the result with 64 at the end of each iteration. So I think the final `len` when we exit this `L_by16_loop_unroll)` loop will be >= 0, right?

BTW: Could you please avoid use `count` as scratch register when say compare `len` with sizes like step_16/64. It's better to uses others like temp0-3.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/18382#discussion_r1677341596