[Bug target/121498] RISC-V: The ra register is not pushed onto the stack due to the use of the -fshrink-wrap parameter

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121498

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2025-08-11
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed. Though the code is not enough to reproduce the issue.

```
(define_insn "*branch<mode>"
  [(set (pc)
        (if_then_else
         (match_operator 1 "ordered_comparison_operator"
                         [(match_operand:X 2 "register_operand" "r")
                          (match_operand:X 3 "reg_or_0_operand" "rJ")])
         (label_ref (match_operand 0 "" ""))
         (pc)))]
  "!TARGET_XCVBI"
{
  if (get_attr_length (insn) == 12)
    return "b%r1\t%2,%z3,1f; jump\t%l0,ra; 1:";

  return "b%C1\t%2,%z3,%l0";
}
  [(set_attr "type" "branch")
   (set_attr "mode" "none")])
```
Basically the problem is the long conditional branch requires ra to do the jump
in some cases but the rtl pattern does not model this at all.