Copy Link
Add to Bookmark
Report

chksum64: how to calculate the ROM checksum for Nintendo 64 ROM images

Nintendo64's profile picture
Published in 
N64 various
 · 30 Oct 2019
Nintendo 64 cartridge.
Pin it
Nintendo 64 cartridge.

chksum64 V1.2 - 16/July/1997


A program for DOS and Linux with included ANSI C source to calculate the ROM checksum for Nintendo64 ROM images. Checksum code reverse engineered from Nagra's program. New in version 1.2 is the feature to checksum files shorter than 1052672 Byte (4K + 1M) for use with the most recent 1.33 BIOS for the V64.

INTRODUCTION


Part of Nintendo 64 ROM images is checksummed, to be precise it is the first MByte after the 4096 Byte header that is checksummed. The N64 verifies the checksum on the ROM when turned on, if it is incorrect it will not run the image. Chksum64 calculates the checksum so that you can run selfwritten programs or patched games on your N64.

Credit must go to Nagra/ Blackbag who originally reverse engineered the checksum algorithm, if he had not done it, it might still be a mistery to us today. Thanks for all the great stuff you have done for the N64 without even owning one!

PROGRAM USAGE


Chksum64 is invoked from the commandline using:

 
chksum64 [-r] [-o|-s] <File>


  • The only required argument is File, the ROM image to be checksummed. It has to be at least 4096 Bytes long, if it is shorter than 4096 + 1 MB the filesize has to be a multiple of four.

  • The -r option is used to turn on read-only mode. In read-only mode the checksum is calculated and displayed on the screen, but it is not written back to the file. This is mainly useful for testing. Read-only mode is also automatically activated when the file cannot be opened for writing.

  • The -o and -s options are used to specify the byte order of the file, in case auto-detection fails. Original byte order is set using -o, this is the way ROM images appear to the N64 internally and also the format produced by compilers. The Doctor V64 uses a different format that can be activated using -s, even and odd bytes are swapped. As mentioned, normally you should not need to use these options because the auto detection will determine the byte order automatically in almost all cases.

SOURCE CODE AVAILABILITY


You should have received the source code file, chksum64.c, with this program, if not feel free to contact me (for my email, see the end of this file). Chksum64 is released under the terms of the GNU public license, its ANSI C source is available to everyone at no charge. The program has been designed to work correctly on platforms with different word sizes and endianness. Limited testing was done on DOS, Linux, Solaris, Irix, and FreeBSD. Use Chksum64 on whatever system you like, use the source in your own programs, but please respect the GNU license. It is there for a reason!

WARNING


If anybody has still got the very first version 1.0 of Chksum64, please be aware the DOS executable distributed with it is infected with the Major.1644 virus. If you used it you should better scan your harddisk for the virus. Sorry for the trouble! (Users of 1.1 need not worry, it is not affected by this problem)

GREETINGS


Greetings to all the people who made the N64 scene what it is. Well, it isn't really anything yet, but finally this is beginning to change. Greetings go out especially to (in no order): Nagra, Stumble, Datawiz, Super_Doc, Sispeo, Widget, Twinsen, RIP, Silo, Fractal, Jovis and all the others from #n64dev, MGallo, of course Hartec and my other friends from Germany, and last but not least Unit and my pals Mike and Ayatollah from Austria.

AUTHORS CONTACT ADDRESS


Send comments or whatever to stan@sbox.tu-graz.ac.at . Flames will be joyfully ignored.

COMPILING


If you are on a Unix system and have gcc in your path, just hit 'make' and you should be done. Otherwise you are forced to figure it out yourself.

NOTE: If you get a bunch of weird error messages, make sure that the Makefile and the source file are in Unix text format and do not have DOS carriage returns at the end of each line.

Makefile


 
CC = gcc
CCFLAGS = -ansi -pedantic -Wall -Wshadow -O2
RM = rm -f

all: chksum64

clean:
$(RM) chksum64

chksum64: chksum64.c
$(CC) $(CCFLAGS) -o chksum64 chksum64.c

chksum64.c


 
/*
chksum64 V1.2, a program to calculate the ROM checksum of Nintendo64 ROMs.
Copyright (C) 1997 Andreas Sterbenz (stan@sbox.tu-graz.ac.at)

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
*/


#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#define max2(a, b) ( (a)>(b) ? (a) : (b) )
#define min2(a, b) ( (a)<(b) ? (a) : (b) )

#define BUFSIZE 32768

#define CHECKSUM_START 0x1000
#define CHECKSUM_LENGTH 0x100000L
#define CHECKSUM_HEADERPOS 0x10
#define CHECKSUM_END (CHECKSUM_START + CHECKSUM_LENGTH)

#define CHECKSUM_STARTVALUE 0xf8ca4ddc

#define ROL(i, b) (((i)<<(b)) | ((i)>>(32-(b))))

#define BYTES2LONG(b, s) ( (((b)[0^(s)] & 0xffL) << 24) | \
(((b)[1^(s)] & 0xffL) << 16) | \
(((b)[2^(s)] & 0xffL) << 8) | \
(((b)[3^(s)] & 0xffL)) )

#define LONG2BYTES(l, b, s) (b)[0^(s)] = ((l)>>24)&0xff; \
(b)[1^(s)] = ((l)>>16)&0xff; \
(b)[2^(s)] = ((l)>> 8)&0xff; \
(b)[3^(s)] = ((l) )&0xff;

#define HEADER_MAGIC 0x80371240

void usage(char *progname)
{
fprintf(stderr, "Usage: %s [-r] [-o|-s] <File>\n\n", progname);
fprintf(stderr, "This program calculates the ROM checksum for Nintendo64 ROM images.\n");
fprintf(stderr, "Checksum code reverse engineered from Nagra's program.\n");
exit(2);
}

int main(int argc, char *argv[])
{
FILE *file1;
char *fname1=NULL, *progname=argv[0];
unsigned char buffer1[BUFSIZE];
long flen1;
unsigned long sum1, sum2;
int swapped=-1;
int readonly = 0;

printf("CHKSUM64 V1.2 Copyright (C) 1997 Andreas Sterbenz (stan@sbox.tu-graz.ac.at)\n");
printf("This program is released under the terms of the GNU public license. NO WARRANTY\n\n");

{
int i;

for( i=1; i<argc; i++ ) {
if( argv[i][0] == '-' ) {
char *c = argv[i]+1;
if( *(c+1) != '\0' ) usage(progname);
switch( *c ) {
case 'r':
readonly = 1;
break;
case 'o':
swapped = 0;
break;
case 's':
swapped = 1;
break;
default:
usage(progname);
}
} else {
if( fname1 == NULL ) {
fname1 = argv[i];
} else {
usage(progname);
}
}
}
}
if( fname1 == NULL ) usage(progname);

if( (readonly == 1) || (file1 = fopen(fname1, "r+b")) == NULL ) {
if( (file1 = fopen(fname1, "rb")) == NULL ) {
fprintf(stderr, "%s: Could not open file '%s' for reading.\n", progname, fname1);
exit(1);
} else {
readonly = 1;
}
}
if( fread(buffer1, 1, 12, file1) != 12 ) {
fprintf(stderr, "%s: File is too short to be a N64 ROM Image, cannot checksum it.\n", progname);
exit(1);
}
if( swapped == -1 ) {
if( BYTES2LONG(buffer1, 0) == HEADER_MAGIC ) {
swapped = 0;
} else if( BYTES2LONG(buffer1, 1) == HEADER_MAGIC ) {
swapped = 1;
}
}
if( swapped != -1 ) {
printf("The image '%s' is in %s format.\n", fname1, (swapped==0) ? "original (not swapped)" : "V64 (byte-swapped)");
} else {
if( (buffer1[8] == 0x80) && (buffer1[9] != 0x80) ) {
swapped = 0;
} else {
if( (buffer1[8] != 0x80) && (buffer1[9] == 0x80) ) {
swapped = 1;
}
}
if( swapped != -1 ) {
printf("WARNING: I am not 100%% certain, but the image '%s' appears to be in %s format.\n", fname1, (swapped==0) ? "original (not swapped)" : "V64 (byte-swapped)");
} else {
fprintf(stderr, "%s: Could not detect type of image '%s', use commandline to override.\n", progname, fname1);
exit(1);
}
}
fseek(file1, 0L, SEEK_END);
flen1 = ftell(file1);
if( flen1 < CHECKSUM_END ) {
if( flen1 < CHECKSUM_START ) {
fprintf(stderr, "%s: File is too short to be a N64 ROM Image, cannot checksum it.\n", progname);
exit(1);
}
if( (flen1 & 3) != 0 ) {
fprintf(stderr, "%s: File length is not a multiple of four, cannot calculate checksum.\n", progname);
exit(1);
}
printf("File is only %ld bytes long, remaining %ld to be checksummed will be assumed 0.\n", flen1, CHECKSUM_END - flen1);
}
fseek(file1, CHECKSUM_START, SEEK_SET);
{
unsigned long i;
unsigned long c1, k1, k2;
unsigned long t1, t2, t3, t4;
unsigned long t5, t6;
unsigned int n;
long clen = CHECKSUM_LENGTH;
long rlen = flen1 - CHECKSUM_START;

/* Below is the actual checksum calculation algorithm, which was
reverse engineered out of Nagra's program.

As you can see, the algorithm is total crap. Obviously it was
designed to be difficult to guess and reverse engineer, and not
to give a good checksum. A simple XOR and ROL 13 would give a
just as good checksum. The ifs and the data dependent ROL are really
extreme nonsense.
*/


t1 = CHECKSUM_STARTVALUE;
t2 = CHECKSUM_STARTVALUE;
t3 = CHECKSUM_STARTVALUE;
t4 = CHECKSUM_STARTVALUE;
t5 = CHECKSUM_STARTVALUE;
t6 = CHECKSUM_STARTVALUE;

for( ;; ) {
if( rlen > 0 ) {
n = fread(buffer1, 1, min2(BUFSIZE, clen), file1);
if( (n & 0x03) != 0 ) {
n += fread(buffer1+n, 1, 4-(n&3), file1);
}
} else {
n = min2(BUFSIZE, clen);
}
if( (n == 0) || ((n&3) != 0) ) {
if( (clen != 0) || (n != 0) ) {
fprintf(stderr, "WARNING: Short read, checksum may be incorrect.\n");
}
break;
}
for( i=0; i<n; i+=4 ) {
c1 = BYTES2LONG(&buffer1[i], swapped);
k1 = t6 + c1;
if( k1 < t6 ) t4++;
t6 = k1;
t3 ^= c1;
k2 = c1 & 0x1f;
k1 = ROL(c1, k2);
t5 += k1;
if( c1 < t2 ) {
t2 ^= k1;
} else {
t2 ^= t6 ^ c1;
}
t1 += c1 ^ t5;
}
if( rlen > 0 ) {
rlen -= n;
if( rlen <= 0 ) memset(buffer1, 0, BUFSIZE);
}
clen -= n;
}
sum1 = t6 ^ t4 ^ t3;
sum2 = t5 ^ t2 ^ t1;
}
LONG2BYTES(sum1, &buffer1[0], 0);
LONG2BYTES(sum2, &buffer1[4], 0);
fseek(file1, CHECKSUM_HEADERPOS, SEEK_SET);
fread(buffer1+8, 1, 8, file1);
printf("Old Checksum: ");
printf("%02X %02X %02X %02X ", buffer1[ 8^swapped], buffer1[ 9^swapped], buffer1[10^swapped], buffer1[11^swapped]);
printf("%02X %02X %02X %02X\n", buffer1[12^swapped], buffer1[13^swapped], buffer1[14^swapped], buffer1[15^swapped]);
printf("Calculated Checksum: ");
printf("%02X %02X %02X %02X ", buffer1[0], buffer1[1], buffer1[2], buffer1[3]);
printf("%02X %02X %02X %02X\n", buffer1[4], buffer1[5], buffer1[6], buffer1[7]);
if( readonly == 0 ) {
fseek(file1, CHECKSUM_HEADERPOS, SEEK_SET);
LONG2BYTES(sum1, &buffer1[16], swapped);
LONG2BYTES(sum2, &buffer1[20], swapped);
if( fwrite(buffer1+16, 1, 8, file1) != 8 ) {
fprintf(stderr, "%s: Could not write new checksum to file '%s'!\n", progname, fname1);
exit(1);
} else {
printf("New checksum successfully written.\n");
}
} else {
printf("File opened in read-only mode, new checksum not written.\n");
}
fclose(file1);
return 0;
}

← previous
next →
loading
sending ...
New to Neperos ? Sign Up for free
download Neperos App from Google Play
install Neperos as PWA

Let's discover also

Recent Articles

Recent Comments

Neperos cookies
This website uses cookies to store your preferences and improve the service. Cookies authorization will allow me and / or my partners to process personal data such as browsing behaviour.

By pressing OK you agree to the Terms of Service and acknowledge the Privacy Policy

By pressing REJECT you will be able to continue to use Neperos (like read articles or write comments) but some important cookies will not be set. This may affect certain features and functions of the platform.
OK
REJECT