mirror of
https://github.com/mii443/wasmer.git
synced 2025-08-25 17:59:27 +00:00
Make trunc_sat() a member of LLVMFunctionCodeGenerator.
This commit is contained in:
Nick Lewycky
@ -569,129 +569,146 @@ fn splat_vector<'ctx>(
|
||||
)
|
||||
}
|
||||
|
||||
// Convert floating point vector to integer and saturate when out of range.
|
||||
// https://github.com/WebAssembly/nontrapping-float-to-int-conversions/blob/master/proposals/nontrapping-float-to-int-conversion/Overview.md
|
||||
fn trunc_sat<'ctx, T: FloatMathType<'ctx>>(
|
||||
builder: &Builder<'ctx>,
|
||||
intrinsics: &Intrinsics<'ctx>,
|
||||
fvec_ty: T,
|
||||
ivec_ty: T::MathConvType,
|
||||
lower_bound: u64, // Exclusive (lowest representable value)
|
||||
upper_bound: u64, // Exclusive (greatest representable value)
|
||||
int_min_value: u64,
|
||||
int_max_value: u64,
|
||||
value: IntValue<'ctx>,
|
||||
name: &str,
|
||||
) -> IntValue<'ctx> {
|
||||
// a) Compare vector with itself to identify NaN lanes.
|
||||
// b) Compare vector with splat of inttofp(upper_bound) to identify
|
||||
// lanes that need to saturate to max.
|
||||
// c) Compare vector with splat of inttofp(lower_bound) to identify
|
||||
// lanes that need to saturate to min.
|
||||
// d) Use vector select (not shuffle) to pick from either the
|
||||
// splat vector or the input vector depending on whether the
|
||||
// comparison indicates that we have an unrepresentable value. Replace
|
||||
// unrepresentable values with zero.
|
||||
// e) Now that the value is safe, fpto[su]i it.
|
||||
// f) Use our previous comparison results to replace certain zeros with
|
||||
// int_min or int_max.
|
||||
impl<'ctx, 'a> LLVMFunctionCodeGenerator<'ctx, 'a> {
|
||||
// Convert floating point vector to integer and saturate when out of range.
|
||||
// https://github.com/WebAssembly/nontrapping-float-to-int-conversions/blob/master/proposals/nontrapping-float-to-int-conversion/Overview.md
|
||||
fn trunc_sat<T: FloatMathType<'ctx>>(
|
||||
&self,
|
||||
fvec_ty: T,
|
||||
ivec_ty: T::MathConvType,
|
||||
lower_bound: u64, // Exclusive (lowest representable value)
|
||||
upper_bound: u64, // Exclusive (greatest representable value)
|
||||
int_min_value: u64,
|
||||
int_max_value: u64,
|
||||
value: IntValue<'ctx>,
|
||||
name: &str,
|
||||
) -> IntValue<'ctx> {
|
||||
// a) Compare vector with itself to identify NaN lanes.
|
||||
// b) Compare vector with splat of inttofp(upper_bound) to identify
|
||||
// lanes that need to saturate to max.
|
||||
// c) Compare vector with splat of inttofp(lower_bound) to identify
|
||||
// lanes that need to saturate to min.
|
||||
// d) Use vector select (not shuffle) to pick from either the
|
||||
// splat vector or the input vector depending on whether the
|
||||
// comparison indicates that we have an unrepresentable value. Replace
|
||||
// unrepresentable values with zero.
|
||||
// e) Now that the value is safe, fpto[su]i it.
|
||||
// f) Use our previous comparison results to replace certain zeros with
|
||||
// int_min or int_max.
|
||||
|
||||
let fvec_ty = fvec_ty.as_basic_type_enum().into_vector_type();
|
||||
let ivec_ty = ivec_ty.as_basic_type_enum().into_vector_type();
|
||||
let fvec_element_ty = fvec_ty.get_element_type().into_float_type();
|
||||
let ivec_element_ty = ivec_ty.get_element_type().into_int_type();
|
||||
let fvec_ty = fvec_ty.as_basic_type_enum().into_vector_type();
|
||||
let ivec_ty = ivec_ty.as_basic_type_enum().into_vector_type();
|
||||
let fvec_element_ty = fvec_ty.get_element_type().into_float_type();
|
||||
let ivec_element_ty = ivec_ty.get_element_type().into_int_type();
|
||||
|
||||
let is_signed = int_min_value != 0;
|
||||
let int_min_value = splat_vector(
|
||||
builder,
|
||||
intrinsics,
|
||||
ivec_element_ty
|
||||
.const_int(int_min_value, is_signed)
|
||||
.as_basic_value_enum(),
|
||||
ivec_ty,
|
||||
"",
|
||||
);
|
||||
let int_max_value = splat_vector(
|
||||
builder,
|
||||
intrinsics,
|
||||
ivec_element_ty
|
||||
.const_int(int_max_value, is_signed)
|
||||
.as_basic_value_enum(),
|
||||
ivec_ty,
|
||||
"",
|
||||
);
|
||||
let lower_bound = if is_signed {
|
||||
builder.build_signed_int_to_float(
|
||||
ivec_element_ty.const_int(lower_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
let is_signed = int_min_value != 0;
|
||||
let int_min_value = splat_vector(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
ivec_element_ty
|
||||
.const_int(int_min_value, is_signed)
|
||||
.as_basic_value_enum(),
|
||||
ivec_ty,
|
||||
"",
|
||||
)
|
||||
} else {
|
||||
builder.build_unsigned_int_to_float(
|
||||
ivec_element_ty.const_int(lower_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
);
|
||||
let int_max_value = splat_vector(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
ivec_element_ty
|
||||
.const_int(int_max_value, is_signed)
|
||||
.as_basic_value_enum(),
|
||||
ivec_ty,
|
||||
"",
|
||||
)
|
||||
};
|
||||
let upper_bound = if is_signed {
|
||||
builder.build_signed_int_to_float(
|
||||
ivec_element_ty.const_int(upper_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
"",
|
||||
)
|
||||
} else {
|
||||
builder.build_unsigned_int_to_float(
|
||||
ivec_element_ty.const_int(upper_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
"",
|
||||
)
|
||||
};
|
||||
);
|
||||
let lower_bound = if is_signed {
|
||||
self.builder.build_signed_int_to_float(
|
||||
ivec_element_ty.const_int(lower_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
"",
|
||||
)
|
||||
} else {
|
||||
self.builder.build_unsigned_int_to_float(
|
||||
ivec_element_ty.const_int(lower_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
"",
|
||||
)
|
||||
};
|
||||
let upper_bound = if is_signed {
|
||||
self.builder.build_signed_int_to_float(
|
||||
ivec_element_ty.const_int(upper_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
"",
|
||||
)
|
||||
} else {
|
||||
self.builder.build_unsigned_int_to_float(
|
||||
ivec_element_ty.const_int(upper_bound, is_signed),
|
||||
fvec_element_ty,
|
||||
"",
|
||||
)
|
||||
};
|
||||
|
||||
let value = builder
|
||||
.build_bitcast(value, fvec_ty, "")
|
||||
.into_vector_value();
|
||||
let zero = fvec_ty.const_zero();
|
||||
let lower_bound = splat_vector(
|
||||
builder,
|
||||
intrinsics,
|
||||
lower_bound.as_basic_value_enum(),
|
||||
fvec_ty,
|
||||
"",
|
||||
);
|
||||
let upper_bound = splat_vector(
|
||||
builder,
|
||||
intrinsics,
|
||||
upper_bound.as_basic_value_enum(),
|
||||
fvec_ty,
|
||||
"",
|
||||
);
|
||||
let nan_cmp = builder.build_float_compare(FloatPredicate::UNO, value, zero, "nan");
|
||||
let above_upper_bound_cmp =
|
||||
builder.build_float_compare(FloatPredicate::OGT, value, upper_bound, "above_upper_bound");
|
||||
let below_lower_bound_cmp =
|
||||
builder.build_float_compare(FloatPredicate::OLT, value, lower_bound, "below_lower_bound");
|
||||
let not_representable = builder.build_or(
|
||||
builder.build_or(nan_cmp, above_upper_bound_cmp, ""),
|
||||
below_lower_bound_cmp,
|
||||
"not_representable_as_int",
|
||||
);
|
||||
let value = builder
|
||||
.build_select(not_representable, zero, value, "safe_to_convert")
|
||||
.into_vector_value();
|
||||
let value = if is_signed {
|
||||
builder.build_float_to_signed_int(value, ivec_ty, "as_int")
|
||||
} else {
|
||||
builder.build_float_to_unsigned_int(value, ivec_ty, "as_int")
|
||||
};
|
||||
let value = builder
|
||||
.build_select(above_upper_bound_cmp, int_max_value, value, "")
|
||||
.into_vector_value();
|
||||
let res = builder
|
||||
.build_select(below_lower_bound_cmp, int_min_value, value, name)
|
||||
.into_vector_value();
|
||||
builder
|
||||
.build_bitcast(res, intrinsics.i128_ty, "")
|
||||
.into_int_value()
|
||||
let value = self
|
||||
.builder
|
||||
.build_bitcast(value, fvec_ty, "")
|
||||
.into_vector_value();
|
||||
let zero = fvec_ty.const_zero();
|
||||
let lower_bound = splat_vector(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
lower_bound.as_basic_value_enum(),
|
||||
fvec_ty,
|
||||
"",
|
||||
);
|
||||
let upper_bound = splat_vector(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
upper_bound.as_basic_value_enum(),
|
||||
fvec_ty,
|
||||
"",
|
||||
);
|
||||
let nan_cmp = self
|
||||
.builder
|
||||
.build_float_compare(FloatPredicate::UNO, value, zero, "nan");
|
||||
let above_upper_bound_cmp = self.builder.build_float_compare(
|
||||
FloatPredicate::OGT,
|
||||
value,
|
||||
upper_bound,
|
||||
"above_upper_bound",
|
||||
);
|
||||
let below_lower_bound_cmp = self.builder.build_float_compare(
|
||||
FloatPredicate::OLT,
|
||||
value,
|
||||
lower_bound,
|
||||
"below_lower_bound",
|
||||
);
|
||||
let not_representable = self.builder.build_or(
|
||||
self.builder.build_or(nan_cmp, above_upper_bound_cmp, ""),
|
||||
below_lower_bound_cmp,
|
||||
"not_representable_as_int",
|
||||
);
|
||||
let value = self
|
||||
.builder
|
||||
.build_select(not_representable, zero, value, "safe_to_convert")
|
||||
.into_vector_value();
|
||||
let value = if is_signed {
|
||||
self.builder
|
||||
.build_float_to_signed_int(value, ivec_ty, "as_int")
|
||||
} else {
|
||||
self.builder
|
||||
.build_float_to_unsigned_int(value, ivec_ty, "as_int")
|
||||
};
|
||||
let value = self
|
||||
.builder
|
||||
.build_select(above_upper_bound_cmp, int_max_value, value, "")
|
||||
.into_vector_value();
|
||||
let res = self
|
||||
.builder
|
||||
.build_select(below_lower_bound_cmp, int_min_value, value, name)
|
||||
.into_vector_value();
|
||||
self.builder
|
||||
.build_bitcast(res, self.intrinsics.i128_ty, "")
|
||||
.into_int_value()
|
||||
}
|
||||
}
|
||||
|
||||
// Convert floating point vector to integer and saturate when out of range.
|
||||
@ -5297,9 +5314,7 @@ impl<'ctx, 'a> LLVMFunctionCodeGenerator<'ctx, 'a> {
|
||||
let (v, i) = self.state.pop1_extra()?;
|
||||
let v = apply_pending_canonicalization(&self.builder, self.intrinsics, v, i);
|
||||
let v = v.into_int_value();
|
||||
let res = trunc_sat(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
let res = self.trunc_sat(
|
||||
self.intrinsics.f32x4_ty,
|
||||
self.intrinsics.i32x4_ty,
|
||||
-2147480000i32 as u32 as u64,
|
||||
@ -5315,9 +5330,7 @@ impl<'ctx, 'a> LLVMFunctionCodeGenerator<'ctx, 'a> {
|
||||
let (v, i) = self.state.pop1_extra()?;
|
||||
let v = apply_pending_canonicalization(&self.builder, self.intrinsics, v, i);
|
||||
let v = v.into_int_value();
|
||||
let res = trunc_sat(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
let res = self.trunc_sat(
|
||||
self.intrinsics.f32x4_ty,
|
||||
self.intrinsics.i32x4_ty,
|
||||
0,
|
||||
@ -5333,9 +5346,7 @@ impl<'ctx, 'a> LLVMFunctionCodeGenerator<'ctx, 'a> {
|
||||
let (v, i) = self.state.pop1_extra()?;
|
||||
let v = apply_pending_canonicalization(&self.builder, self.intrinsics, v, i);
|
||||
let v = v.into_int_value();
|
||||
let res = trunc_sat(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
let res = self.trunc_sat(
|
||||
self.intrinsics.f64x2_ty,
|
||||
self.intrinsics.i64x2_ty,
|
||||
std::i64::MIN as u64,
|
||||
@ -5351,9 +5362,7 @@ impl<'ctx, 'a> LLVMFunctionCodeGenerator<'ctx, 'a> {
|
||||
let (v, i) = self.state.pop1_extra()?;
|
||||
let v = apply_pending_canonicalization(&self.builder, self.intrinsics, v, i);
|
||||
let v = v.into_int_value();
|
||||
let res = trunc_sat(
|
||||
&self.builder,
|
||||
self.intrinsics,
|
||||
let res = self.trunc_sat(
|
||||
self.intrinsics.f64x2_ty,
|
||||
self.intrinsics.i64x2_ty,
|
||||
std::u64::MIN,
|
||||
|
||||
Reference in New Issue
Block a user