Make remaining float intrinsics require float arguments (#57398)

The `fptrunc`/`fpext` intrinsics were modified in #57160 to throw on
non-float arguments.
- The arithmetic and math float intrinsics now require all their
arguments to be floats
- `fptosi`/`fptoui` require their source to be a float
- `sitofp`/`uitofp` require their destination type to be a float

Also fixes #57384.

Author: xal-0

Compiler/src/tfuncs.jl

@@ -2432,6 +2432,43 @@ const _SPECIAL_BUILTINS = Any[
     Core._apply_iterate,
 ]
 
+# Intrinsics that require all arguments to be floats
+const _FLOAT_INTRINSICS = Any[
+    Intrinsics.neg_float,
+    Intrinsics.add_float,
+    Intrinsics.sub_float,
+    Intrinsics.mul_float,
+    Intrinsics.div_float,
+    Intrinsics.min_float,
+    Intrinsics.max_float,
+    Intrinsics.fma_float,
+    Intrinsics.muladd_float,
+    Intrinsics.neg_float_fast,
+    Intrinsics.add_float_fast,
+    Intrinsics.sub_float_fast,
+    Intrinsics.mul_float_fast,
+    Intrinsics.div_float_fast,
+    Intrinsics.min_float_fast,
+    Intrinsics.max_float_fast,
+    Intrinsics.eq_float,
+    Intrinsics.ne_float,
+    Intrinsics.lt_float,
+    Intrinsics.le_float,
+    Intrinsics.eq_float_fast,
+    Intrinsics.ne_float_fast,
+    Intrinsics.lt_float_fast,
+    Intrinsics.le_float_fast,
+    Intrinsics.fpiseq,
+    Intrinsics.abs_float,
+    Intrinsics.copysign_float,
+    Intrinsics.ceil_llvm,
+    Intrinsics.floor_llvm,
+    Intrinsics.trunc_llvm,
+    Intrinsics.rint_llvm,
+    Intrinsics.sqrt_llvm,
+    Intrinsics.sqrt_llvm_fast
+]
+
 # Types compatible with fpext/fptrunc
 const CORE_FLOAT_TYPES = Union{Core.BFloat16, Float16, Float32, Float64}
 
@@ -2849,7 +2886,8 @@ function intrinsic_exct(𝕃::AbstractLattice, f::IntrinsicFunction, argtypes::V
             return ErrorException
         end
 
-        # fpext and fptrunc have further restrictions on the allowed types.
+        # fpext, fptrunc, fptoui, fptosi, uitofp, and sitofp have further
+        # restrictions on the allowed types.
         if f === Intrinsics.fpext &&
             !(ty <: CORE_FLOAT_TYPES && xty <: CORE_FLOAT_TYPES && Core.sizeof(ty) > Core.sizeof(xty))
             return ErrorException
@@ -2858,6 +2896,12 @@ function intrinsic_exct(𝕃::AbstractLattice, f::IntrinsicFunction, argtypes::V
             !(ty <: CORE_FLOAT_TYPES && xty <: CORE_FLOAT_TYPES && Core.sizeof(ty) < Core.sizeof(xty))
             return ErrorException
         end
+        if (f === Intrinsics.fptoui || f === Intrinsics.fptosi) && !(xty <: CORE_FLOAT_TYPES)
+            return ErrorException
+        end
+        if (f === Intrinsics.uitofp || f === Intrinsics.sitofp) && !(ty <: CORE_FLOAT_TYPES)
+            return ErrorException
+        end
 
         return Union{}
     end
@@ -2870,11 +2914,15 @@ function intrinsic_exct(𝕃::AbstractLattice, f::IntrinsicFunction, argtypes::V
         return Union{}
     end
 
-    # The remaining intrinsics are math/bits/comparison intrinsics. They work on all
-    # primitive types of the same type.
+    # The remaining intrinsics are math/bits/comparison intrinsics.
+    # All the non-floating point intrinsics work on primitive values of the same type.
     isshift = f === shl_int || f === lshr_int || f === ashr_int
     argtype1 = widenconst(argtypes[1])
     isprimitivetype(argtype1) || return ErrorException
+    if contains_is(_FLOAT_INTRINSICS, f)
+        argtype1 <: CORE_FLOAT_TYPES || return ErrorException
+    end
+
     for i = 2:length(argtypes)
         argtype = widenconst(argtypes[i])
         if isshift ? !isprimitivetype(argtype) : argtype !== argtype1

Compiler/test/effects.jl

@@ -1401,3 +1401,29 @@ end == Compiler.EFFECTS_UNKNOWN
 @test !Compiler.intrinsic_nothrow(Core.Intrinsics.fpext, Any[Type{Float32}, Float64])
 @test !Compiler.intrinsic_nothrow(Core.Intrinsics.fpext, Any[Type{Int32}, Float16])
 @test !Compiler.intrinsic_nothrow(Core.Intrinsics.fpext, Any[Type{Float32}, Int16])
+
+# Float intrinsics require float arguments
+@test Base.infer_effects((Int16,)) do x
+    return Core.Intrinsics.abs_float(x)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((Int32, Int32)) do x, y
+    return Core.Intrinsics.add_float(x, y)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((Int32, Int32)) do x, y
+    return Core.Intrinsics.add_float(x, y)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((Int64, Int64, Int64)) do x, y, z
+    return Core.Intrinsics.fma_float(x, y, z)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((Int64,)) do x
+    return Core.Intrinsics.fptoui(UInt32, x)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((Int64,)) do x
+    return Core.Intrinsics.fptosi(Int32, x)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((Int64,)) do x
+    return Core.Intrinsics.sitofp(Int64, x)
+end |> !Compiler.is_nothrow
+@test Base.infer_effects((UInt64,)) do x
+    return Core.Intrinsics.uitofp(Int64, x)
+end |> !Compiler.is_nothrow

src/APInt-C.cpp

@@ -321,7 +321,7 @@ void LLVMFPtoInt(jl_datatype_t *ty, void *pa, jl_datatype_t *oty, integerPart *p
         Val = julia_bfloat_to_float(*(uint16_t*)pa);
     else if (ty == jl_float32_type)
         Val = *(float*)pa;
-    else if (jl_float64_type)
+    else if (ty == jl_float64_type)
         Val = *(double*)pa;
     else
         jl_error("FPtoSI: runtime floating point intrinsics are not implemented for bit sizes other than 16, 32 and 64");
@@ -352,7 +352,7 @@ void LLVMFPtoInt(jl_datatype_t *ty, void *pa, jl_datatype_t *oty, integerPart *p
     else {
         APFloat a(Val);
         bool isVeryExact;
-        APFloat::roundingMode rounding_mode = APFloat::rmNearestTiesToEven;
+        APFloat::roundingMode rounding_mode = RoundingMode::TowardZero;
         unsigned nbytes = alignTo(onumbits, integerPartWidth) / host_char_bit;
         integerPart *parts = (integerPart*)alloca(nbytes);
         APFloat::opStatus status = a.convertToInteger(MutableArrayRef<integerPart>(parts, nbytes), onumbits, isSigned, rounding_mode, &isVeryExact);

src/intrinsics.cpp

@@ -676,17 +676,23 @@ static jl_cgval_t generic_cast(
     Type *to = bitstype_to_llvm((jl_value_t*)jlto, ctx.builder.getContext(), true);
     Type *vt = bitstype_to_llvm(v.typ, ctx.builder.getContext(), true);
 
-    // fptrunc fpext depend on the specific floating point format to work
-    // correctly, and so do not pun their argument types.
+    // fptrunc and fpext depend on the specific floating point
+    // format to work correctly, and so do not pun their argument types.
     if (!(f == fpext || f == fptrunc)) {
-        if (toint)
-            to = INTT(to, DL);
-        else
-            to = FLOATT(to);
-        if (fromint)
-            vt = INTT(vt, DL);
-        else
-            vt = FLOATT(vt);
+        // uitofp/sitofp require a specific float type argument
+        if (!(f == uitofp || f == sitofp)){
+            if (toint)
+                to = INTT(to, DL);
+            else
+                to = FLOATT(to);
+        }
+        // fptoui/fptosi require a specific float value argument
+        if (!(f == fptoui || f == fptosi)) {
+            if (fromint)
+                vt = INTT(vt, DL);
+            else
+                vt = FLOATT(vt);
+        }
     }
 
     if (!to || !vt)
@@ -1428,10 +1434,13 @@ static jl_cgval_t emit_intrinsic(jl_codectx_t &ctx, intrinsic f, jl_value_t **ar
         if (!jl_is_primitivetype(xinfo.typ))
             return emit_runtime_call(ctx, f, argv, nargs);
         Type *xtyp = bitstype_to_llvm(xinfo.typ, ctx.builder.getContext(), true);
-        if (float_func()[f])
-            xtyp = FLOATT(xtyp);
-        else
+        if (float_func()[f]) {
+            if (!xtyp->isFloatingPointTy())
+                return emit_runtime_call(ctx, f, argv, nargs);
+        }
+        else {
             xtyp = INTT(xtyp, DL);
+        }
         if (!xtyp)
             return emit_runtime_call(ctx, f, argv, nargs);
         ////Bool are required to be in the range [0,1]

src/runtime_intrinsics.c

@@ -1073,31 +1073,26 @@ typedef void (fintrinsic_op1)(unsigned, jl_value_t*, void*, void*);
 static inline jl_value_t *jl_fintrinsic_1(jl_value_t *ty, jl_value_t *a, const char *name, fintrinsic_op1 *bfloatop, fintrinsic_op1 *halfop, fintrinsic_op1 *floatop, fintrinsic_op1 *doubleop)
 {
     jl_task_t *ct = jl_current_task;
-    if (!jl_is_primitivetype(jl_typeof(a)))
+    jl_datatype_t *aty = (jl_datatype_t *)jl_typeof(a);
+    if (!jl_is_primitivetype(aty))
         jl_errorf("%s: value is not a primitive type", name);
     if (!jl_is_primitivetype(ty))
         jl_errorf("%s: type is not a primitive type", name);
     unsigned sz2 = jl_datatype_size(ty);
     jl_value_t *newv = jl_gc_alloc(ct->ptls, sz2, ty);
     void *pa = jl_data_ptr(a), *pr = jl_data_ptr(newv);
-    unsigned sz = jl_datatype_size(jl_typeof(a));
-    switch (sz) {
-    /* choose the right size c-type operation based on the input */
-    case 2:
-        if (jl_typeof(a) == (jl_value_t*)jl_float16_type)
-            halfop(sz2 * host_char_bit, ty, pa, pr);
-        else /*if (jl_typeof(a) == (jl_value_t*)jl_bfloat16_type)*/
-            bfloatop(sz2 * host_char_bit, ty, pa, pr);
-        break;
-    case 4:
+
+    if (aty == jl_float16_type)
+        halfop(sz2 * host_char_bit, ty, pa, pr);
+    else if (aty == jl_bfloat16_type)
+        bfloatop(sz2 * host_char_bit, ty, pa, pr);
+    else if (aty == jl_float32_type)
         floatop(sz2 * host_char_bit, ty, pa, pr);
-        break;
-    case 8:
+    else if (aty == jl_float64_type)
         doubleop(sz2 * host_char_bit, ty, pa, pr);
-        break;
-    default:
-        jl_errorf("%s: runtime floating point intrinsics are not implemented for bit sizes other than 16, 32 and 64", name);
-    }
+    else
+        jl_errorf("%s: runtime floating point intrinsics require both arguments to be Float16, BFloat16, Float32, or Float64", name);
+
     return newv;
 }
 
@@ -1273,30 +1268,24 @@ JL_DLLEXPORT jl_value_t *jl_##name(jl_value_t *a, jl_value_t *b) \
 { \
     jl_task_t *ct = jl_current_task; \
     jl_value_t *ty = jl_typeof(a); \
+    jl_datatype_t *aty = (jl_datatype_t *)ty; \
     if (jl_typeof(b) != ty) \
         jl_error(#name ": types of a and b must match"); \
     if (!jl_is_primitivetype(ty)) \
         jl_error(#name ": values are not primitive types"); \
     int sz = jl_datatype_size(ty); \
     jl_value_t *newv = jl_gc_alloc(ct->ptls, sz, ty); \
     void *pa = jl_data_ptr(a), *pb = jl_data_ptr(b), *pr = jl_data_ptr(newv); \
-    switch (sz) { \
-    /* choose the right size c-type operation */ \
-    case 2: \
-        if ((jl_datatype_t*)ty == jl_float16_type) \
-            jl_##name##16(16, pa, pb, pr); \
-        else /*if ((jl_datatype_t*)ty == jl_bfloat16_type)*/ \
-            jl_##name##bf16(16, pa, pb, pr); \
-        break; \
-    case 4: \
+    if (aty == jl_float16_type) \
+        jl_##name##16(16, pa, pb, pr); \
+    else if (aty == jl_bfloat16_type) \
+        jl_##name##bf16(16, pa, pb, pr); \
+    else if (aty == jl_float32_type) \
         jl_##name##32(32, pa, pb, pr); \
-        break; \
-    case 8: \
+    else if (aty == jl_float64_type) \
         jl_##name##64(64, pa, pb, pr); \
-        break; \
-    default: \
-        jl_error(#name ": runtime floating point intrinsics are not implemented for bit sizes other than 16, 32 and 64"); \
-    } \
+    else \
+        jl_error(#name ": runtime floating point intrinsics require both arguments to be Float16, BFloat16, Float32, or Float64"); \
     return newv; \
 }
 
@@ -1308,30 +1297,24 @@ JL_DLLEXPORT jl_value_t *jl_##name(jl_value_t *a, jl_value_t *b) \
 JL_DLLEXPORT jl_value_t *jl_##name(jl_value_t *a, jl_value_t *b) \
 { \
     jl_value_t *ty = jl_typeof(a); \
+    jl_datatype_t *aty = (jl_datatype_t *)ty; \
     if (jl_typeof(b) != ty) \
         jl_error(#name ": types of a and b must match"); \
     if (!jl_is_primitivetype(ty)) \
         jl_error(#name ": values are not primitive types"); \
     void *pa = jl_data_ptr(a), *pb = jl_data_ptr(b); \
-    int sz = jl_datatype_size(ty); \
     int cmp; \
-    switch (sz) { \
-    /* choose the right size c-type operation */ \
-    case 2: \
-        if ((jl_datatype_t*)ty == jl_float16_type) \
-            cmp = jl_##name##16(16, pa, pb); \
-        else /*if ((jl_datatype_t*)ty == jl_bfloat16_type)*/ \
-            cmp = jl_##name##bf16(16, pa, pb); \
-        break; \
-    case 4: \
+    if (aty == jl_float16_type) \
+        cmp = jl_##name##16(16, pa, pb); \
+    else if (aty == jl_bfloat16_type) \
+        cmp = jl_##name##bf16(16, pa, pb); \
+    else if (aty == jl_float32_type) \
         cmp = jl_##name##32(32, pa, pb); \
-        break; \
-    case 8: \
+    else if (aty == jl_float64_type) \
         cmp = jl_##name##64(64, pa, pb); \
-        break; \
-    default: \
-        jl_error(#name ": runtime floating point intrinsics are not implemented for bit sizes other than 32 and 64"); \
-    } \
+    else \
+        jl_error(#name ": runtime floating point intrinsics require both arguments to be Float16, BFloat16, Float32, or Float64"); \
+ \
     return cmp ? jl_true : jl_false; \
 }
 
@@ -1344,30 +1327,24 @@ JL_DLLEXPORT jl_value_t *jl_##name(jl_value_t *a, jl_value_t *b, jl_value_t *c)
 { \
     jl_task_t *ct = jl_current_task; \
     jl_value_t *ty = jl_typeof(a); \
+    jl_datatype_t *aty = (jl_datatype_t *)ty; \
     if (jl_typeof(b) != ty || jl_typeof(c) != ty) \
         jl_error(#name ": types of a, b, and c must match"); \
     if (!jl_is_primitivetype(ty)) \
         jl_error(#name ": values are not primitive types"); \
     int sz = jl_datatype_size(ty); \
     jl_value_t *newv = jl_gc_alloc(ct->ptls, sz, ty); \
     void *pa = jl_data_ptr(a), *pb = jl_data_ptr(b), *pc = jl_data_ptr(c), *pr = jl_data_ptr(newv); \
-    switch (sz) { \
-    /* choose the right size c-type operation */ \
-    case 2: \
-        if ((jl_datatype_t*)ty == jl_float16_type) \
+    if (aty == jl_float16_type) \
             jl_##name##16(16, pa, pb, pc, pr); \
-        else /*if ((jl_datatype_t*)ty == jl_bfloat16_type)*/ \
+    else if (aty == jl_bfloat16_type) \
             jl_##name##bf16(16, pa, pb, pc, pr); \
-        break; \
-    case 4: \
+    else if (aty == jl_float32_type) \
         jl_##name##32(32, pa, pb, pc, pr); \
-        break; \
-    case 8: \
+    else if (aty == jl_float64_type) \
         jl_##name##64(64, pa, pb, pc, pr); \
-        break; \
-    default: \
-        jl_error(#name ": runtime floating point intrinsics are not implemented for bit sizes other than 16, 32 and 64"); \
-    } \
+    else \
+        jl_error(#name ": runtime floating point intrinsics require both arguments to be Float16, BFloat16, Float32, or Float64"); \
     return newv; \
 }
 
@@ -1661,7 +1638,7 @@ static inline void fptrunc(jl_datatype_t *aty, void *pa, jl_datatype_t *ty, void
     fptrunc_convert(float64, bfloat16);
     fptrunc_convert(float64, float32);
     else
-        jl_error("fptrunc: runtime floating point intrinsics are not implemented for bit sizes other than 16, 32 and 64");
+        jl_error("fptrunc: runtime floating point intrinsics require both arguments to be Float16, BFloat16, Float32, or Float64");
 #undef fptrunc_convert
 }
 
@@ -1685,7 +1662,7 @@ static inline void fpext(jl_datatype_t *aty, void *pa, jl_datatype_t *ty, void *
     fpext_convert(bfloat16, float64);
     fpext_convert(float32, float64);
     else
-        jl_error("fptrunc: runtime floating point intrinsics are not implemented for bit sizes other than 16, 32 and 64");
+        jl_error("fptrunc: runtime floating point intrinsics require both arguments to be Float16, BFloat16, Float32, or Float64");
 #undef fpext_convert
 }