FLP37-C. Cast the return value of a function that returns a floating point type

C99 Section 6.8.6.4 3 states

If a return statement with an expression is executed, the value of the expression is returned to the caller as the value of the function call expression. If the expression has a type different from the return type of the function in which it appears, the value is converted as if by assignment to an object having the return type of the function. 142)

This paragraph is footnoted as follows:

142) The return statement is not an assignment. The overlap restriction of subclause 6.5.16.1 does not apply to the case of function return. The representation of floating-point values may have wider range or precision. A cast may be used to remove this extra range and precision.

Conversion as if by assignment to the type of the function is required if the return expression has a different type than the function, but not if the return expression has a wider value only because of wide evaluation. This allows seemingly inconsistent and confusing behavior. Consider,

float f(float x) { return x * 0.1f; }

float g(float x) { return x * 0.1; }

Function f is allowed to return a value wider than float, but function g (which uses the wider constant) is not.

Although the current text does not require narrowing return expressions of the same type as the function, it does not clearly state what is allowed. Is it allowed to narrow the result? Is it allowed to narrow the result sometimes but not always? Is it allowed to partially narrow the result (e.g., if the ABI returns floats in double format but a float function has a float return expression evaluated to wider than double)? An aggressive implementation could argue “yes” to all these, though the resulting behavior would complicate debugging and error analysis.

Footnote 142 says a cast may be used to remove extra range and precision from the return expression. This means a predictable program must have casts on all floating-point function calls (except where the function directly feeds an operator like assignment that implies the conversion). With type-generic math (tgmath), the programmer has to reason through the tgmath resolution rules to determine which casts to apply. These are significant obstacles to writing predictable code.

NOTE: WG14 has voted to include the following text in C1X. This only impacts implementations that implement the optional Annex F "IEC 60559 floating-point arithmetic".

Require return expressions to be converted as if by assignment to the type of the function, but only in Annex F. This is a compromise that addresses the problems for Annex F implementations while not impacting non-Annex F implementations that exercise the license for wide returns.

Insert the following new subclause after F.5 (and increment subsequent subclause numbers):

F.6 The return statement

If the return expression is evaluated in a floating-point format different from the return type, then the expression is converted to the return type of the function and the resulting value is returned to the caller.

Noncompliant Code Example

The following example code has been constructed to illustrate an example that does not conform to this recommendation.  The code is non-conforming because it does not cast the result of the expression in the return statement and thereby guarantee the range or precision is no wider than expected.  The uncertainty in this example is introduced by the constant 0.1.  This constant may be stored with a range or precision that is greater than that of float.  Consequently, the result of x * 0.1 may also have a range or precision greater than that of float.  As described above, this range or precision may not be reduced to that of a float and, thus, the caller of g() may have a value that is more precise than expected.  This may lead to inconsistent program execution across different platforms.

float g(float x) {
  return x * 0.1;
}

Compliant Code Example

The following code example remedies the above noncompliant code by casting the value of the expression in the return statement.  This forces the return value to have the expected range and precision as described in Section 5.2.4.2.2 9 of the C Standard.

float g(float x) {
  return (float)(x * 0.1);
}

Risk Assessment

Failure to follow this guideline can lead to inconsistent results across different platforms.

Related Vulnerabilities

Search for vulnerabilities resulting from the violation of this rule on the CERT website.

Other Languages

References