[Bug target/120689] [12/13/14/15/16 Regression] Codegen optimization regression passing struct in register in gcc 10+ on x86-64 since r10-577

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120689

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So, can't we just keep cris doing what it wants and not penalize other arches
because of that?
Like:
--- gcc/function.cc.jj  2025-05-20 08:14:06.105410349 +0200
+++ gcc/function.cc     2025-06-17 20:03:03.457413807 +0200
@@ -2937,7 +2937,7 @@ assign_parm_setup_block (struct assign_p
   if (stack_parm == 0)
     {
       HOST_WIDE_INT parm_align
-       = (STRICT_ALIGNMENT
+       = ((STRICT_ALIGNMENT || BITS_PER_WORD <= MAX_SUPPORTED_STACK_ALIGNMENT)
           ? MAX (DECL_ALIGN (parm), BITS_PER_WORD) : DECL_ALIGN (parm));

       SET_DECL_ALIGN (parm, parm_align);