https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119966
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
validate_subreg allows the paradoxical subreg even in the case of hard
register:
```
/* Paradoxical subregs must have offset zero. */
if (maybe_gt (osize, isize))
return known_eq (offset, 0U);
/* This is a normal subreg. Verify that the offset is representable. */
/* For hard registers, we already have most of these rules collected in
subreg_offset_representable_p. */
if (reg && REG_P (reg) && HARD_REGISTER_P (reg))
...
So either validate_subreg needs to reject if the subreg of a hard reg is not
valid for that mode for paradoxical subreg. Or I am using
rtl_hooks.gen_lowpart_no_emit incorrectly in my patch to create a paradoxical
subreg.
