Skip to end of metadata
Go to start of metadata

Bit-fields can be used to allow flags or other integer values with small ranges to be packed together to save storage space.

It is implementation-defined whether the specifier int designates the same type as signed int or the same type as unsigned int for bit-fields. According to the C Standard [ISO/IEC 9899:2011], C integer promotions also require that "if an int can represent all values of the original type (as restricted by the width, for a bit-field), the value is converted to an int; otherwise, it is converted to an unsigned int."

This issue is similar to the signedness of plain char, discussed in INT07-C. Use only explicitly signed or unsigned char type for numeric values. A plain int bit-field that is treated as unsigned will promote to int as long as its field width is less than that of int because int can hold all values of the original type. This behavior is the same as that of a plain char treated as unsigned. However, a plain int bit-field treated as unsigned will promote to unsigned int if its field width is the same as that of int. This difference makes a plain int bit-field even trickier than a plain char.

Bit-field types other than _Bool, int, signed int, and unsigned int are implementation-defined. They still obey the integer promotions quoted previously when the specified width is at least as narrow as CHAR_BIT*sizeof(int), but wider bit-fields are not portable.

Noncompliant Code Example

This noncompliant code depends on implementation-defined behavior. It prints either -1 or 255, depending on whether a plain int bit-field is signed or unsigned.

struct {
  int a: 8;
} bits = {255};

int main(void) {
  printf("bits.a = %d.\n", bits.a);
  return 0;
}

Compliant Solution

This compliant solution uses an unsigned int bit-field and does not depend on implementation-defined behavior:

struct {
  unsigned int a: 8;
} bits = {255};

int main(void) {
  printf("bits.a = %d.\n", bits.a);
  return 0;
}

Risk Assessment

Making invalid assumptions about the type of a bit-field or its layout can result in unexpected program flow.

Recommendation

Severity

Likelihood

Remediation Cost

Priority

Level

INT12-C

Low

Unlikely

Medium

P2

L3

Automated Detection

Tool

Version

Checker

Description

Astrée
19.04
bitfield-typeFully checked
Axivion Bauhaus Suite

6.9.0

CertC-INT12
CodeSonar
5.0p0
LANG.TYPE.BFSIGNBit-field signedness not explicit
Compass/ROSE




ECLAIR

1.2

CC2.INT12

Fully implemented

Klocwork
2018
MISRA.BITFIELD.TYPE
LDRA tool suite
9.7.1

73 S

Fully implemented

Parasoft C/C++test
10.4.2

CERT_C-INT12-a

Bit fields shall only be defined to be of type unsigned int or signed int

Polyspace Bug Finder

R2018a

Integer conversion overflow

MISRA C:2012 Rule 6.1

Overflow when converting between integer types

Bit-fields shall only be declared with an appropriate type

PRQA QA-C
9.5

0634 (I)

0635

Fully implemented
RuleChecker
19.04
bitfield-typeFully checked
SonarQube C/C++ Plugin
3.11
S814

Related Vulnerabilities

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

Related Guidelines

Bibliography

[ISO/IEC 9899:2011]Subclause 6.3.1.1, "Boolean, Characters, and Integers"



3 Comments

  1. Is this the correct interpretation of the implementation-defined behavior?

    I always thought the issue was similar to that associated with char, signed char and unsigned char. i.e. if a bit-field is of type int, then it is implementation defined if it is signed or unsigned. If the signedness is specified, the behavior (when used in an expression) is then appropriate for the type of the definition. e.g.

    struct {            int m : 8;   } a;
    struct {   unsigned int m : 8;   } ua;
    struct {     signed int m : 8;   } sa;
    
    
     
    

    In the above, a.m has implementation defined signedness, ua.m is unsigned and sa.m is signed.

    This is the interpretation that MISRA uses. 

  2. The real issue is almost the same as that for plain char.  This rule originally contradicted itself by quoting the integer promotions in the second paragraph, and then later saying that it was implementation-defined whether an 8-bit unsigned int bit-field was promoted to int or unsigned int.  The integer promotions are the correct interpretation.  This is now fixed.

  3. If it's signed int, the 255 initializer is invalid and should generate a diagnostic.

    I don't have a good replacement, since my personal rule is "don't use bit fields, ever".