1 #ifndef GSIMM_H
2 #define GSIMM_H
4 /* Length of file message digest (MD) in bytes. Longer MD's are
5 better, but increase processing time for diminishing returns.
6 Must be multiple of NUM_HASHES_PER_CHAR / 8, and at least 24
7 for good results
8 */
9 #define MD_LENGTH 32
10 #define MD_BITS (MD_LENGTH * 8)
12 /* The MIN_FILE_SIZE indicates the absolute minimal file size that
13 can be processed. As indicated above, the first and last
14 RABIN_WINDOW_SIZE - 1 bytes are skipped.
15 In order to get at least an average of 12 samples
16 per bit in the final message digest, require at least 3 * MD_LENGTH
17 complete windows in the file. */
18 #define MIN_FILE_SIZE (3 * MD_LENGTH + 2 * (RABIN_WINDOW_SIZE - 1))
20 /* Limit matching algorithm to files less than 256 MB, so we can use
21 32 bit integers everywhere without fear of overflow. For larger
22 files we should add logic to mmap the file by piece and accumulate
23 the frequency counts. */
24 #define MAX_FILE_SIZE (256*1024*1024 - 1)
26 void gb_simm_process(u_char *data, unsigned len, u_char *md);
28 #endif