Auto merge of rust-lang#121728 - tgross35:f16-f128-step1-ty-updates, r=compiler-errors

Add stubs in IR and ABI for `f16` and `f128`

This is the very first step toward the changes in rust-lang#114607 and the [`f16` and `f128` RFC](https://rust-lang.github.io/rfcs/3453-f16-and-f128.html). It adds the types to `rustc_type_ir::FloatTy` and `rustc_abi::Primitive`, and just propagates those out as `unimplemented!` stubs where necessary.

These types do not parse yet so there is no feature gate, and it should be okay to use `unimplemented!`.

The next steps will probably be AST support with parsing and the feature gate.

r? `@compiler-errors`
cc `@Nilstrieb` suggested breaking the PR up in rust-lang#120645 (comment)

Author: bors

clippy_lints/src/float_literal.rs

@@ -81,6 +81,7 @@ impl<'tcx> LateLintPass<'tcx> for FloatLiteral {
                 LitFloatType::Unsuffixed => None,
             };
             let (is_whole, is_inf, mut float_str) = match fty {
+                FloatTy::F16 => unimplemented!("f16_f128"),
                 FloatTy::F32 => {
                     let value = sym_str.parse::<f32>().unwrap();
 
@@ -91,6 +92,7 @@ impl<'tcx> LateLintPass<'tcx> for FloatLiteral {
 
                     (value.fract() == 0.0, value.is_infinite(), formatter.format(value))
                 },
+                FloatTy::F128 => unimplemented!("f16_f128"),
             };
 
             if is_inf {
@@ -135,8 +137,10 @@ impl<'tcx> LateLintPass<'tcx> for FloatLiteral {
 #[must_use]
 fn max_digits(fty: FloatTy) -> u32 {
     match fty {
+        FloatTy::F16 => unimplemented!("f16_f128"),
         FloatTy::F32 => f32::DIGITS,
         FloatTy::F64 => f64::DIGITS,
+        FloatTy::F128 => unimplemented!("f16_f128"),
     }
 }
 

clippy_utils/src/consts.rs

@@ -778,8 +778,10 @@ pub fn mir_to_const<'tcx>(lcx: &LateContext<'tcx>, result: mir::Const<'tcx>) ->
                 let range = alloc_range(offset + size * idx, size);
                 let val = alloc.read_scalar(&lcx.tcx, range, /* read_provenance */ false).ok()?;
                 res.push(match flt {
+                    FloatTy::F16 => unimplemented!("f16_f128"),
                     FloatTy::F32 => Constant::F32(f32::from_bits(val.to_u32().ok()?)),
                     FloatTy::F64 => Constant::F64(f64::from_bits(val.to_u64().ok()?)),
+                    FloatTy::F128 => unimplemented!("f16_f128"),
                 });
             }
             Some(Constant::Vec(res))