Some operators do not evaluate their operands beyond the type information the operands provide. When using one of these operators, do not pass an operand that would otherwise yield a side effect since the side effect will not be generated.

The sizeof operator yields the size (in bytes) of its operand, which may be an expression or the parenthesized name of a type.  In most cases, the operand is not evaluated.  A possible exception is when the type of the operand is a variable length array type (VLA); then the expression is evaluated. When part of the operand of the sizeof operator is a VLA type and when changing the value of the VLA's size expression would not affect the result of the operator, it is unspecified whether or not the size expression is evaluated. (See unspecified behavior 22.)

The operand passed to_Alignof is never evaluated, despite not being an expression. For instance, if the operand is a VLA type and the VLA's size expression contains a side effect, that side effect is never evaluated.

The operand used in the controlling expression of a _Generic selection expression is never evaluated.

Providing an expression that appears to produce side effects may be misleading to programmers who are not aware that these expressions are not evaluated, and in the case of a VLA used in sizeof, have unspecified results. As a result, programmers may make invalid assumptions about program state, leading to errors and possible software vulnerabilities.

This rule is similar to PRE31-C. Avoid side effects in arguments to unsafe macros.

Noncompliant Code Example (sizeof)

In this noncompliant code example, the expression a++ is not evaluated:

#include <stdio.h>
 
void func(void) {
  int a = 14;
  int b = sizeof(a++);
  printf("%d, %d\n", a, b);
}

Consequently, the value of a after b has been initialized is 14.

Compliant Solution (sizeof)

In this compliant solution, the variable a is incremented outside of the sizeof operation:

#include <stdio.h>
 
void func(void) {
  int a = 14;
  int b = sizeof(a);
  ++a;
  printf("%d, %d\n", a, b);
}

Noncompliant Code Example (sizeof, VLA)

In this noncompliant code example, the expression ++n in the initialization expression of a must be evaluated because its value affects the size of the VLA operand of the sizeof operator. However, in the initialization expression of b, the expression ++n % 1 evaluates to 0. This means that the value of n does not affect the result of the sizeof operator. Consequently, it is unspecified whether or not n will be incremented when initializing b.

#include <stddef.h>
#include <stdio.h>
  
void f(size_t n) {
  /* n must be incremented */ 
  size_t a = sizeof(int[++n]);
 
  /* n need not be incremented */
  size_t b = sizeof(int[++n % 1 + 1]);

  printf("%zu, %zu, %zu\n", a, b, n);
  /* ... */
}

Compliant Solution (sizeof, VLA)

This compliant solution avoids changing the value of the variable n used in each sizeof expression and instead increments n safely afterwards:

#include <stddef.h>
#include <stdio.h>
  
void f(size_t n) {
  size_t a = sizeof(int[n + 1]);
  ++n;

  size_t b = sizeof(int[n % 1 + 1]);
  ++n;
  printf("%zu, %zu, %zu\n", a, b, n);
  /* ... */
}

Noncompliant Code Example (_Generic)

This noncompliant code example attempts to modify a variable's value as part of the _Generic selection control expression. The programmer may expect that a is incremented, but because _Generic does not evaluate its control expression, the value of a is not modified.

#include <stdio.h>

#define S(val) _Generic(val, int : 2, \
                             short : 3, \
                             default : 1)
void func(void) {
  int a = 0;
  int b = S(a++);
  printf("%d, %d\n", a, b);
}

Compliant Solution (_Generic)

In this compliant solution, a is incremented outside of the _Generic selection expression:

#include <stdio.h>

#define S(val) _Generic(val, int : 2, \
                             short : 3, \
                             default : 1)
void func(void) {
  int a = 0;
  int b = S(a);
  ++a;
  printf("%d, %d\n", a, b);
} 

Noncompliant Code Example (_Alignof)

This noncompliant code example attempts to modify a variable while getting its default alignment value. The user may have expected val to be incremented as part of the _Alignof expression, but because _Alignof does not evaluate its operand, val is unchanged.

#include <stdio.h>
 
void func(void) {
  int val = 0; 
  /* ... */ 
  size_t align = _Alignof(int[++val]);
  printf("%zu, %d\n", align, val);
  /* ... */
}

Compliant Solution (_Alignof)

 This compliant solution moves the expression out of the _Alignof operator:

#include <stdio.h>
void func(void) {
  int val = 0; 
  /* ... */ 
  ++val;
  size_t align = _Alignof(int[val]);
  printf("%zu, %d\n", align, val);
  /* ... */
}

Exceptions

EXP44-C-EX1: Reading a volatile-qualified value is a side-effecting operation. However, accessing a value through a volatile-qualified type does not guarantee side effects will happen on the read of the value unless the underlying object is also volatile-qualified. Idiomatic reads of a volatile-qualified object are permissible as an operand to a sizeof()_Alignof(), or _Generic expression, as in the following example:

void f(void) {
  int * volatile v;
  (void)sizeof(*v);
}

Risk Assessment

If expressions that appear to produce side effects are supplied to an operator that does not evaluate its operands, the results may be different than expected. Depending on how this result is used, it can lead to unintended program behavior.

Rule

Severity

Likelihood

Remediation Cost

Priority

Level

EXP44-C

Low

Unlikely

Low

P3

L3

Automated Detection

Tool

Version

Checker

Description

Astrée
24.04

alignof-side-effect
generic-selection-side-effect
sizeof

Fully checked
Axivion Bauhaus Suite

7.2.0

CertC-EXP44
Clang
3.9
-Wunevaluated-expressionCan diagnose some instance of this rule, but not all (such as the _Alignof NCCE).
CodeSonar
8.1p0
LANG.STRUCT.SE.SIZEOF
LANG.STRUCT.SE.CGEN
Side effects in sizeof
Side Effects in C Generic Selection
Compass/ROSE




Coverity
2017.07

MISRA C 2004 Rule 12.3

Partially implemented

ECLAIR

1.2

CC2.EXP06

Fully implemented

Helix QAC

2024.3

C3307
Klocwork
2024.3
MISRA.SIZEOF.SIDE_EFFECT
LDRA tool suite
9.7.1

54 S, 653 S

Fully implemented

Parasoft C/C++test
2023.1

CERT_C-EXP44-a
CERT_C-EXP44-b

Object designated by a volatile lvalue should not be accessed in the operand of the sizeof operator
The function call that causes the side effect shall not be the operand of the sizeof operator

PC-lint Plus

1.4

9006

Partially supported: reports use of sizeof with an expression that would have side effects

Polyspace Bug Finder

R2024a

CERT C: Rule EXP44-C


Checks for situations when side effects of specified expressions are ignored (rule fully covered)

PVS-Studio

7.33

V568
RuleChecker
24.04

alignof-side-effect
generic-selection-side-effect
sizeof

Fully checked

Related Vulnerabilities

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

Related Guidelines

Key here (explains table format and definitions)

Taxonomy

Taxonomy item

Relationship

CERT CEXP52-CPP. Do not rely on side effects in unevaluated operandsPrior to 2018-01-12: CERT: Unspecified Relationship





13 Comments

  1. This is an example of a bug, since nobody would intentionally code a side effect in that context.  Indeed it would be nice for compilers to warn about side-effects in sizeof operands.

  2. dumb question.  most of the examples in the standard which take the sizeof a unary expression and not a type don't use parenthesis.  Should we follow suite?

    1. Personally, I don't like that style. I find it to be confusing and inconsistent (and potentially error-prone if you don't remember that only a unary expression is valid).

       

  3. The string argument in the call to printf in Compliant Solution (_Alignof) is missing a closing right quote.

  4. I am concerned that the risk assessment values are incorrect. The severity is currently Low, indicating DOS or hang. And the probability is Unlikely (the lowest setting). I am thinking that perhaps the severity should be high, and liklihood at least Probable (medium setting).

    Here is code example that has a buffer overflow b/c of violating this rule:

    int main(void) {
      char src[100];
      size_t size = sizeof(src-1);  // oops should be sizeof(str) - 1.
      // Expected: 99, got 8 (on ILP64)
      char *dest = calloc(size, sizeof(char));
      if (dest == NULL) {
        abort(); /* handle error */
      }
      printf("What is your name?\n");
      fgets( src, sizeof(src), stdin);
      strcpy(dest, src); // buffer overflow!
      printf("Hello, %s\n", dest);
      free(dest);
      return 0;
    }
    1. Will Klieber says:

      In your example, I don't see any side effects in the operand of sizeof.  (And I don't see any use of  _Alignof or _Generic.)  Am I missing something?

      1. I'm with Will on this one – I don't think the code you posted violates this rule as written (nor do I think the rule should be modified to cover this case).

        1. So which rule covers (or should cover) my code example?  (There are analysis tools that flag any arithmetic expression in a sizeof operation.)


          1. I'd go with ARR38-C ("Guarantee that library functions do not form invalid pointers").

            1. That would catch my code example, but what if I had said sizeof(src + 1)?  That would have complied with ARR38-C, and still produced a buffer overflow in my code.

              1. I think both your original and revised examples are covered by MEM35-C (Allocate sufficient memory for an object).

    2. I think your "oops" comment should also use "src" not "str" as well.  But, I don't think we have any rules on misleading comments.  (wink)