Crypto++  8.5
Free C++ class library of cryptographic schemes
Macros | Typedefs | Variables
config_int.h File Reference

Library configuration file. More...

Go to the source code of this file.

Macros

#define CRYPTOPP_NO_GLOBAL_BYTE   1
 Library byte guard. More...
 
#define W64LIT(x)   ...
 Declare an unsigned word64. More...
 
#define SW64LIT(x)   ...
 Declare a signed word64. More...
 
#define CRYPTOPP_BOOL_SLOW_WORD64   ...
 Declare ops on word64 are slow. More...
 
#define CRYPTOPP_WORD128_AVAILABLE   ...
 128-bit word availability More...
 

Typedefs

typedef unsigned char byte
 8-bit unsigned datatype More...
 
typedef unsigned short word16
 16-bit unsigned datatype More...
 
typedef unsigned int word32
 32-bit unsigned datatype More...
 
typedef signed char sbyte
 8-bit signed datatype More...
 
typedef signed short sword16
 16-bit signed datatype More...
 
typedef signed int sword32
 32-bit signed datatype More...
 
typedef unsigned long long word64
 64-bit unsigned datatype More...
 
typedef signed long long sword64
 64-bit signed datatype More...
 
typedef __uint128_t word128
 128-bit unsigned datatype More...
 
typedef word64 lword
 Large word type. More...
 
typedef word32 hword
 Half word used for multiprecision integer arithmetic. More...
 
typedef word64 word
 Full word used for multiprecision integer arithmetic. More...
 
typedef word128 dword
 Double word used for multiprecision integer arithmetic. More...
 

Variables

const lword LWORD_MAX = ...
 Large word type max value. More...
 
const unsigned int WORD_SIZE = sizeof(word)
 Size of a platform word in bytes. More...
 
const unsigned int WORD_BITS = WORD_SIZE * 8
 Size of a platform word in bits. More...
 

Detailed Description

Library configuration file.

config_int.h provides defines and typedefs for fixed size integers. The library's choices for fixed size integers predates other standard-based integers by about 5 years. After fixed sizes were made standard, the library continued to use its own definitions for compatibility with previous versions of the library.

config.h was split into components in May 2019 to better integrate with Autoconf and its feature tests. The splitting occurred so users could continue to include config.h while allowing Autoconf to write new config_asm.h and new config_cxx.h using its feature tests.

Note
You should include config.h rather than config_int.h directly.
See also
Issue 835, Make config.h more autoconf friendly, Configure.sh script on the Crypto++ wiki
Since
Crypto++ 8.3

Definition in file config_int.h.

Macro Definition Documentation

◆ CRYPTOPP_NO_GLOBAL_BYTE

#define CRYPTOPP_NO_GLOBAL_BYTE   1

Library byte guard.

CRYPTOPP_NO_GLOBAL_BYTE indicates byte is in the Crypto++ namespace.

The Crypto++ byte was originally in global namespace to avoid ambiguity with other byte typedefs. byte was moved to CryptoPP namespace at Crypto++ 6.0 due to C++17, std::byte and potential compile problems.

See also
Issue 442, std::byte on the Crypto++ wiki
Since
Crypto++ 6.0

Definition at line 41 of file config_int.h.

◆ W64LIT

#define W64LIT (   x)    ...

Declare an unsigned word64.

W64LIT is used to portability declare or assign 64-bit literal values. W64LIT will append the proper suffix to ensure the compiler accepts the literal.

Use the macro like shown below.

   word64 x = W64LIT(0xffffffffffffffff);
 
Since
Crypto++ 1.0

Definition at line 119 of file config_int.h.

◆ SW64LIT

#define SW64LIT (   x)    ...

Declare a signed word64.

SW64LIT is used to portability declare or assign 64-bit literal values. SW64LIT will append the proper suffix to ensure the compiler accepts the literal.

Use the macro like shown below.

   sword64 x = SW64LIT(0xffffffffffffffff);
 
Since
Crypto++ 8.0

Definition at line 129 of file config_int.h.

◆ CRYPTOPP_BOOL_SLOW_WORD64

#define CRYPTOPP_BOOL_SLOW_WORD64   ...

Declare ops on word64 are slow.

CRYPTOPP_BOOL_SLOW_WORD64 is typically defined to 1 on platforms that have a machine word smaller than 64-bits. That is, the define is present on 32-bit platforms. The define is also present on platforms where the cpu is slow even with a 64-bit cpu.

Definition at line 136 of file config_int.h.

◆ CRYPTOPP_WORD128_AVAILABLE

#define CRYPTOPP_WORD128_AVAILABLE   ...

128-bit word availability

CRYPTOPP_WORD128_AVAILABLE indicates a 128-bit word is available from the platform. 128-bit words are usually available on 64-bit platforms, but not available 32-bit platforms.

If CRYPTOPP_WORD128_AVAILABLE is not defined, then 128-bit words are not available.

GCC and compatible compilers signal 128-bit word availability with the preporcessor macro __SIZEOF_INT128__ >= 16.

Since
Crypto++ 2.0

Definition at line 204 of file config_int.h.

Typedef Documentation

◆ byte

typedef unsigned char byte

8-bit unsigned datatype

The Crypto++ byte was originally in global namespace to avoid ambiguity with other byte typedefs. byte was moved to CryptoPP namespace at Crypto++ 6.0 due to C++17, std::byte and potential compile problems.

See also
CRYPTOPP_NO_GLOBAL_BYTE, Issue 442, std::byte on the Crypto++ wiki
Since
Crypto++ 1.0, CryptoPP namespace since Crypto++ 6.0

Definition at line 56 of file config_int.h.

◆ word16

typedef unsigned short word16

16-bit unsigned datatype

Since
Crypto++ 1.0

Definition at line 59 of file config_int.h.

◆ word32

typedef unsigned int word32

32-bit unsigned datatype

Since
Crypto++ 1.0

Definition at line 62 of file config_int.h.

◆ sbyte

typedef signed char sbyte

8-bit signed datatype

The 8-bit signed datatype was added to support constant time implementations for curve25519, X25519 key agreement and ed25519 signatures.

Since
Crypto++ 8.0

Definition at line 69 of file config_int.h.

◆ sword16

typedef signed short sword16

16-bit signed datatype

The 32-bit signed datatype was added to support constant time implementations for curve25519, X25519 key agreement and ed25519 signatures.

Since
Crypto++ 8.0

Definition at line 75 of file config_int.h.

◆ sword32

typedef signed int sword32

32-bit signed datatype

The 32-bit signed datatype was added to support constant time implementations for curve25519, X25519 key agreement and ed25519 signatures.

Since
Crypto++ 8.0

Definition at line 81 of file config_int.h.

◆ word64

typedef unsigned long long word64

64-bit unsigned datatype

The typedef for word64 varies depending on the platform. On Microsoft platforms it is unsigned __int64. On Unix & Linux with LP64 data model it is unsigned long. On Unix & Linux with ILP32 data model it is unsigned long long.

Since
Crypto++ 1.0

Definition at line 91 of file config_int.h.

◆ sword64

typedef signed long long sword64

64-bit signed datatype

The typedef for sword64 varies depending on the platform. On Microsoft platforms it is signed __int64. On Unix & Linux with LP64 data model it is signed long. On Unix & Linux with ILP32 data model it is signed long long.

Since
Crypto++ 8.0

Definition at line 99 of file config_int.h.

◆ word128

typedef __uint128_t word128

128-bit unsigned datatype

The typedef for word128 varies depending on the platform. word128 is only available on 64-bit machines when CRYPTOPP_WORD128_AVAILABLE is defined. On Unix & Linux with LP64 data model it is __uint128_t. Microsoft platforms do not provide a 128-bit integer type. 32-bit platforms do not provide a 128-bit integer type.

Since
Crypto++ 5.6

Definition at line 109 of file config_int.h.

◆ lword

typedef word64 lword

Large word type.

lword is a typedef for large word types. It is used for file offsets and such.

Definition at line 158 of file config_int.h.

◆ hword

typedef word32 hword

Half word used for multiprecision integer arithmetic.

hword is used for multiprecision integer arithmetic. The typedef for hword varies depending on the platform. On 32-bit platforms it is usually word16. On 64-bit platforms it is usually word32.

Library users typically use byte, word16, word32 and word64.

Since
Crypto++ 2.0

Definition at line 174 of file config_int.h.

◆ word

typedef word64 word

Full word used for multiprecision integer arithmetic.

word is used for multiprecision integer arithmetic. The typedef for word varies depending on the platform. On 32-bit platforms it is usually word32. On 64-bit platforms it is usually word64.

Library users typically use byte, word16, word32 and word64.

Since
Crypto++ 2.0

Definition at line 182 of file config_int.h.

◆ dword

typedef word128 dword

Double word used for multiprecision integer arithmetic.

dword is used for multiprecision integer arithmetic. The typedef for dword varies depending on the platform. On 32-bit platforms it is usually word64. On 64-bit Unix & Linux platforms it is usually word128. word128 is not available on Microsoft platforms. word128 is only available when CRYPTOPP_WORD128_AVAILABLE is defined.

Library users typically use byte, word16, word32 and word64.

See also
CRYPTOPP_WORD128_AVAILABLE
Since
Crypto++ 2.0

Definition at line 193 of file config_int.h.

Variable Documentation

◆ LWORD_MAX

const lword LWORD_MAX = ...

Large word type max value.

LWORD_MAX is the maximum value for large word types. Since an lword is an unsigned type, the value is 0xffffffffffffffff. W64LIT will append the proper suffix.

Definition at line 164 of file config_int.h.

◆ WORD_SIZE

const unsigned int WORD_SIZE = sizeof(word)

Size of a platform word in bytes.

The size of a platform word, in bytes

Definition at line 245 of file config_int.h.

◆ WORD_BITS

const unsigned int WORD_BITS = WORD_SIZE * 8

Size of a platform word in bits.

The size of a platform word, in bits

Definition at line 249 of file config_int.h.