rename min/maxnum intrinsics to min/maximum_number and fix their LLVM lowering

Author: RalfJung

src/intrinsics/mod.rs

@@ -1266,7 +1266,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             ret.write_cvalue(fx, val);
         }
 
-        sym::minnumf16 => {
+        sym::minimum_number_nsz_f16 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1275,7 +1275,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f16));
             ret.write_cvalue(fx, val);
         }
-        sym::minnumf32 => {
+        sym::minimum_number_nsz_f32 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1284,7 +1284,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f32));
             ret.write_cvalue(fx, val);
         }
-        sym::minnumf64 => {
+        sym::minimum_number_nsz_f64 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1293,7 +1293,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f64));
             ret.write_cvalue(fx, val);
         }
-        sym::minnumf128 => {
+        sym::minimum_number_nsz_f128 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1302,7 +1302,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f128));
             ret.write_cvalue(fx, val);
         }
-        sym::maxnumf16 => {
+        sym::maximum_number_nsz_f16 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1311,7 +1311,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f16));
             ret.write_cvalue(fx, val);
         }
-        sym::maxnumf32 => {
+        sym::maximum_number_nsz_f32 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1320,7 +1320,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f32));
             ret.write_cvalue(fx, val);
         }
-        sym::maxnumf64 => {
+        sym::maximum_number_nsz_f64 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);
@@ -1329,7 +1329,7 @@ fn codegen_regular_intrinsic_call<'tcx>(
             let val = CValue::by_val(val, fx.layout_of(fx.tcx.types.f64));
             ret.write_cvalue(fx, val);
         }
-        sym::maxnumf128 => {
+        sym::maximum_number_nsz_f128 => {
             intrinsic_args!(fx, args => (a, b); intrinsic);
             let a = a.load_scalar(fx);
             let b = b.load_scalar(fx);

src/num.rs

@@ -498,10 +498,10 @@ fn codegen_ptr_binop<'tcx>(
     }
 }
 
-// In Rust floating point min and max don't propagate NaN. In Cranelift they do however.
-// For this reason it is necessary to use `a.is_nan() ? b : (a >= b ? b : a)` for `minnumf*`
-// and `a.is_nan() ? b : (a <= b ? b : a)` for `maxnumf*`. NaN checks are done by comparing
-// a float against itself. Only in case of NaN is it not equal to itself.
+// In Rust floating point min and max don't propagate NaN (not even SNaN). In Cranelift they do
+// however. For this reason it is necessary to use `a.is_nan() ? b : (a >= b ? b : a)` for
+// `minnumf*` and `a.is_nan() ? b : (a <= b ? b : a)` for `maxnumf*`. NaN checks are done by
+// comparing a float against itself. Only in case of NaN is it not equal to itself.
 pub(crate) fn codegen_float_min(fx: &mut FunctionCx<'_, '_, '_>, a: Value, b: Value) -> Value {
     // FIXME(bytecodealliance/wasmtime#8312): Replace with Cranelift `fcmp` once
     // `f16`/`f128` backend lowerings have been added to Cranelift.