On Mon, 17 Mar 2025 14:43:05 GMT, Chen Liang <li...@openjdk.org> wrote:
> It seems this is an overflow for the TC part of the flag. The RC part seems > to be able to overflow too, but because it occupies the higher bits its > overflow will never affect the TC. And this is the only site where TC is > masked with `UMASK` after overflow-capable addition. Yes, from this point RC (in)accurate masking can be considered cosmetic, but this might be the source of this bug with TC. > Should we reexamine all uses of `UMASK` to mask only RC/TC explicitly? '(v.stackPred & SP_MASK) | (UC_MASK & c)' patterns seem ok to me, but for all subfield arithmetic it would be more clear to see all subfields recombined. ------------- PR Comment: https://git.openjdk.org/jdk/pull/24034#issuecomment-2729913449
