[Bug middle-end/123175] [14/15/16 Regression] Wrong folding of VEC_PERM_EXPR

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123175

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rsandifo at gcc dot gnu.org
   Target Milestone|---                         |14.4
            Summary|Wrong folding of            |[14/15/16 Regression] Wrong
                   |VEC_PERM_EXPR               |folding of VEC_PERM_EXPR
           Keywords|                            |wrong-code

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Note there's much uses of 'nelts' in the pattern.  Careful auditing needs to be
done to see whether simply deriving nelts from op0 subparts is correct.

We accept such differently sized inputs since GCC 14, they are mostly
generated by the vectorizer for I think VLA input vectors only(?)