Incorrect tautology diagnostic with native half‑precision

[codemzs]

-Wunreachable-code is silent on targets with native FP16 support

Problem

The tautology/contradiction check in CheckIncorrectLogicOperator fails on targets with native scalar FP16 (e.g., AArch64 +fullfp16). It still works on x86‑64, so the lapse went unnoticed.

Triple	FP16 handling	Warning?
x86_64-unknown-linux-gnu	__fp16 / _Float16 are promoted to float (extra cast)	✔️ emitted
aarch64-unknown-linux-gnu +fullfp16	no promotion for _Float16(native arithmetic)	❌ missing

---

Why promotion hides the bug

static void test(_Float16 x) {
  if (x != 0 || x != 1.0)   // should warn: always true
    x = 0.9;
  else
    x = 0.8;
}

x86‑64 AST — warning appears

BinaryOperator 0x555564fcdfc0 '_Bool' '!='
|-ImplicitCastExpr 0x555564fcdf90 'float' **<FloatingCast>**
| -ImplicitCastExpr 0x555564fcdf78 '_Float16' <LValueToRValue>
|   -DeclRefExpr 0x555564fcdf38 '_Float16' lvalue ParmVar 0x555564fcdd78 'x' '__fp16 &'
-ImplicitCastExpr 0x555564fcdfa8 'float' <IntegralToFloating>
  -IntegerLiteral 0x555564fcdf58 'int' 0

BinaryOperator 0x555564fce050 '_Bool' '!='
|-ImplicitCastExpr 0x555564fce038 'double' **<FloatingCast>**
| -ImplicitCastExpr 0x555564fce020 '_Float16' <LValueToRValue>
|   -DeclRefExpr 0x555564fcdfe0 '_Float16' lvalue ParmVar 0x555564fcdd78 'x' '__fp16 &'
`-FloatingLiteral 0x555564fce000 'double' 1.000000e+00

AArch64 +fullfp16 AST — warning lost

BinaryOperator 0x55556438dbe8 '_Bool' '!='
|-ImplicitCastExpr 0x55556438dbb8 '_Float16' **<LValueToRValue>**
| -DeclRefExpr 0x55556438db78 '_Float16' lvalue ParmVar 0x55556438d9b8 'x' '__fp16 &'
-ImplicitCastExpr 0x55556438dbd0 '_Float16' <IntegralToFloating>
  -IntegerLiteral 0x55556438db98 'int' 0

BinaryOperator 0x55556438dc78 '_Bool' '!='
|-ImplicitCastExpr 0x55556438dc60 'double' **<FloatingCast>**
| -ImplicitCastExpr 0x55556438dc48 '_Float16' <LValueToRValue>
|   -DeclRefExpr 0x55556438dc08 '_Float16' lvalue ParmVar 0x55556438d9b8 'x' '__fp16 &'
-FloatingLiteral 0x55556438dc28 'double' 1.000000e+00

No FloatingCast → operands’ cast chains differ →Expr::isSameComparisonOperand() returns false → diagnostic short‑circuits.

[llvmbot]

@llvm/issue-subscribers-clang-frontend

Author: M. Zeeshan Siddiqui (codemzs)

`-Wunreachable-code` is silent on targets with native FP16 support ----------------------------------------------------------------------------

Problem

The tautology/contradiction check in CheckIncorrectLogicOperator fails on targets with native scalar FP16 (e.g., AArch64 +fullfp16).
It still works on x86‑64, so the lapse went unnoticed.

Triple	FP16 handling	Warning?
x86_64-unknown-linux-gnu	__fp16 / _Float16 are promoted to float (extra cast)	✔️ emitted
aarch64-unknown-linux-gnu +fullfp16	no promotion (native arithmetic)	❌ missing

---

Why promotion hides the bug

x86‑64 AST — warning appears

BinaryOperator '_Bool' '!='
├─ ImplicitCastExpr 'float' <FloatingCast>
│  └─ ImplicitCastExpr '__fp16' <LValueToRValue>
│     └─ DeclRefExpr '__fp16' ParmVar 'x'
└─ ImplicitCastExpr 'float' <IntegralToFloating>
   └─ IntegerLiteral 'int' 0

AArch64 +fullfp16 AST — warning lost

BinaryOperator '_Bool' '!='
├─ ImplicitCastExpr '__fp16' <LValueToRValue>
│  └─ DeclRefExpr '__fp16' ParmVar 'x'
└─ ImplicitCastExpr '__fp16' <IntegralToFloating>
   └─ IntegerLiteral 'int' 0

No FloatingCast → operands’ cast chains differ →Expr::isSameComparisonOperand() returns false → diagnostic short‑circuits.