[Bug target/122266] [13/14/15/16 regression] miscompilation of int128 xor-and-shift under -O2 in combine

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122266

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #7)
> Why not just unsigned char?  Is record_value_for_reg ever called with OImode
> or XImode?

We might want to at least add an assert then that unsigned char is sufficient.
unsigned char would be consistent with

  unsigned char                 sign_bit_copies;