[Bug tree-optimization/125811] `a * (1 << b)` is not transformed into `a << b` in some cases

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125811

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement
           Keywords|                            |easyhack,
                   |                            |missed-optimization
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2026-06-16
             Status|UNCONFIRMED                 |NEW
            Summary|Possible Missed             |`a * (1 << b)` is not
                   |Optimization at O3          |transformed into `a << b`
                   |                            |in some cases

--- Comment #1 from Drea Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.
This pattern:
```
/* Fold (a * (1 << b)) into (a << b)  */
(simplify
 (mult:c @0 (convert? (lshift integer_onep@1 @2)))
  (if (! FLOAT_TYPE_P (type)
       && tree_nop_conversion_p (type, TREE_TYPE (@1)))
   (lshift @0 @2)))

```

Only supports nop conversion But in this case we have a truncating conversion:
```
  _1 = (intD.7) v0_u8_3(D);
  _2 = 1 << _1;
  i0_u8_4 = (uint8_tD.4679) _2;
  i1_u8_5 = v0_u8_3(D) * i0_u8_4;
```

I think we can do the conversion and truncate afterwards.

So it becomes (convert (lshift (convert @0) @2))