Message Digests
File hash.c is the control module for running message digests. It
calls out to specific symmetric encryption routines for the specific algorithms
supported.
Constants
The following constant is defined (in abstract.h)
-
MAXHASHSIZE=32 largest supported digest
size
Data Types
md_context
The Message Digest context structure is private; and its contents are
dependent on the algorithm employed.
Public Methods
All functions within this module are public
-
md_context hashInit(byte md_algor);
-
The chosen message digest algorithm is instantiated; the return value is
a handle on this instance.
-
void hashUpdate(md_context contextArg,
byte *buf, uint32_t count);
-
A byte array buf containing count bytes is added to the (possibly
null) set of bytes already digested in the given md_context..
-
void hashFinal(md_context *contextArg,
byte *digest);
-
The md_context is destroyed, yielding the result of all data submitted.
-
int hashDigest(byte md_algor);
-
Returns the length in bytes of the message digest. The space pointed to
by the digest argument of hashFinal() must be at least this
long.
-
boolean hashAlgAvail(byte md_algor);
-
Indicate whether the byte value indicates a supported message digest algorithm
-
boolean hashAlgRecognised(byte
md_algor);
-
Indicate whether the byte value indicates a recognised message digest algorithm.
This may return TRUE for certain algorithms the hashAlgAvail()
returns FALSE, indicating that algorithm is known but not implemented.
-
Dependencies
Each algorithm supported must implement the following interface, where
the name, or some fragment of the name, of the algorithm replaces <ALG>
An initialiser
void <ALG>Init(void **context, size_t *length);
returning the hash context and its length.
a do-it routine
void <ALG>Update(void *context, byte *buff, uint32_t len);
to add a buffer's worth of new data
a closing routine
void <ALG>Final(void **context, byte *digest, size_t length);
concluding the digesting
and #define constant <ALG>HASHSIZE to give the digest
size in bytes
The digest algorithms
Currently these supporting routines are defined in 3way.c,
haval*.c,
md5.c, rmd160.c and sha.c. Each of these algorithms
has its own peculiar internal workings, and where possible, a set of test
vectors and a test program which can be compiled in stand-alone fashion..
The implementations have all been augmented from their usual forms to support
the hash interface specified above. The 3Way based hash algorithm implements
one of the procedures whereby a block cipher can be run as a message digest,
and provides a way of creating hashed data of the size this algorithm requires
as a key. Both SHA and SHA-1 are provided, with the addition of the extra
bit rotation in SHA-1 being defined by a run-time flag.
The process of adding new algorithms is mechanical.It requires modifications
to several functions in hash.c. In each, an extra
case must
be added to a switch statement to process the new message digest,
if the hash algorithm matchs the appropriate key value.
-
hashInit
-
Add case to call <ALG>Init.
-
hashUpdate
-
Add case to call <ALG>Update.
-
hashFinal
-
Add case to call <ALG>Final.
-
hashDigest
-
Add case to return appropriate resulting hashsize (in bytes).
-
hashAlgAvail
-
Add case to return TRUE, to indicate the cipher is available.
webmaster@bifroest.demon.co.uk
