[Bug middle-end/114169] miss optimization of repeat load&store in place

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114169

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2024-03-07
     Ever confirmed|0                           |1
          Component|rtl-optimization            |middle-end

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.

Reduced testcase:
```
struct S0 {
   int  f0;
   int t[3+4];
};

struct S0 g_2;
struct S0 g_4;

void func_1() {
  //struct S0 b = g_2;
  g_4 = g_2;
  struct S0 b = g_4;
  g_4.f0 = 0;
  g_2 = b;
}

void func_2() {
  struct S0 b = g_2;
  g_4 = b;
  g_4.f0 = 0;
  g_2 = b;
}
```

We still get an extra load/store with func_2 but it least it is not a dependent
load/store ...

Basically GCC doesn't notice that g_2 = b = g_2 was a no-op.