ShuriZma
/
dolphin
mirror of https://github.com/dolphin-emu/dolphin.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
dolphin/docs/ActionReplay/GCNCodeTypes.txt

956 lines
26 KiB
Plaintext
Raw Normal View History

an attempt to move docs to trunk. god tortoisesvn was an ass to do this with tortoisesvn and google code (core pdf was refused multiple times) NOTE: docs/WiiMote/Core V2.1 + EDR.pdf is still missing deu to googlecode hating me git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@3491 8ced0084-cf51-0410-be5f-012b33b47a6e
2009-06-18 16:07:45 +00:00
--------------------------------
GCN AR CODES TYPES EXPLANATION
--------------------------------
v1.1 by kenobi
History :
v1.1 : Removed the 'NCT' codes types.
Added 'Byte Copy', 'Pointer Mod' and 'AND/OR' codes type.
Added some notes about the (m) codes, the 'write to CCxxxxxx' code.
Fixed some typos.
v1.0 : Initial release.
Special thanks to Parasyte for his help/informations about some codes types.
This document has been written for educational purpose.
It may help you create codes for the GCN AR, or they might be useless
junk... Your call !
If you know the GBA's ARv3 codes types, you'll find the GCN AR codes types
quite similar...
Also note that the PS2's AR MAX codes types are very close to the GCN ones.
Warning : This document is meant for advanced codes creators, NOT FOR
NEWBIES OR WANNABES. Sorry.
*************** If you are an experienced, known (by me or gscentral
admins) code hacker, and you don't
understand this document, you may try to ask help using the
www.gscentral.com forums
(or PM me there).
Special Note 1 : All adresses MUST be compatible with the data size you want
the codes are using.
*************** That means : -ANY address can be used for BYTE
reading/writing.
-Address MUST be a multiple of 2 for HALFWORD
reading/writing.
(Last hex number of the address must be either
:0,2,4,6,8,A,C,E)
-Address MUST be a multiple of 4 for WORD
reading/writing.
(Last hex number of the address must be either
:0,4,8,C)
If you don't follow this rule, the codes won't work (or the
AR might crash)...!
Special Note 2 : All codes are formatted like that : XXXXXXXX YYYYYYYY.
I called ADDRESS (in caps) the XXXXXXXX, and VALUE (in
caps) the YYYYYYYY.
Special Note 3 : GCN memory range is 0x80000000 - 0x817FFFFF cached, and
0xC0000000 - 0xC17FFFFF uncached.
(don't ask what it means, I don't get it either :P).
The codes will usualy write to the cached area.
Special Note 4 : The codes type numbers I give after a code name is a number
created like this:
For "Type zX" codes , the number X is :
AAA (3 most significant bits of the code's "VALUE")
For normal codes, the number in parenthesis after the name
of the code is :
AAABBCC (7 most significant bits of the code's
"ADDRESS")
AAA : type bits.
BB : subtype bits.
CC : value bits.
You can use them as reference, or just ignore them...
Special Note 5 : Any "unused" data could be filled with random numbers to
create a "unique encryption",
which could "sign" your codes. I randomly explained how it
works. It might not work
with every code. This feature isn't really interessing, but
I felt like it should be
noticed.
Special Note 6 : "Register 1BB4" is one of the register (= a given place in
the NGC memory) that the AR
uses to store some data while executing codes.
Special Note 7 : The addresses, values, and all the numbers starting by
"0x", or having the letter(s)
A, B, C, D, E and/or F in them are Hexadecimal numbers. If
you don't know what hexadecimal
is, make a search in Google.
Special Note 8 : If you don't know C/C++, be aware that "<<" means "Shift
left", and ">>" "Shift right".
"Shift left" is the "Lsh" button of the Windows calculator
(in "Scientific" mode).
"Shift right" is gotten by clicking the "Inv" checkbox,
then the "Lsh" button of the
Windows calculator (in "Scientific" mode).
________________
----------------
| Type z Codes |
----------------
"Type z" are codes which have an ADDRESS eqal to 00000000 ("z" stands for
"zero").
For any "Type zX" codes : X = code type = (VALUE >> 29) AND 0x07.
(If X>4, the code will be skipped)
-------------------------------
// Type "z0" : END OF CODES //
-------------------------------
1 line code.
00000000 00000000
It means "end of the code" (or "no more codes are executed").
The AR will "give" back the hand to the game, and then will start execute
codes
from the very 1st of the list.
--------------------------------------------
// Type "z2" : Normal execution of codes //
--------------------------------------------
1 line code.
00000000 40000000
Set register 1BB4 to 0.
It means that the AR goes back to the normal execution of codes.
(And it should break a "stop executing codes", set when register 1BB4 is =
2).
-----------------------------------------------------
// Type "z3" : Executes all codes in the same row //
-----------------------------------------------------
1 line code.
00000000 60000000
Set register 1BB4 to 1.
It means the AR will execute all the codes, without giving back the hand to
the
game, unless register 1BB4 changes value (with a "z2" code for exemple).
-------------------------------
// Type "z4" : Fill & Slide //
-------------------------------
2 lines code.
00000000 8XXXXXXX
Y1Y2Y3Y4 Z1Z2Z3Z4
Address = 8XXXXXXX AND 0x81FFFFFF.
Size = (address >> 25) AND 0x03.
(Size 0 = 8bits, Size 1 = 16 bits, Size 2 = 32 bits. Size 3 = Unused)
Value = Y1Y2Y3Y4.
Address increment = 0000Z3Z4 if (Z1 >> 3 = 0).
= FFFFZ3Z4 if (Z1 >> 3 = 1).
NOTE : When using halfword (or word), make address increment >> 1 (or >> 2)
when
computing the code.
Value increment = 00000000Z1 if (Z1 >> 3 = 0).
= FFFFFFFFZ1 if (Z1 >> 3 = 1).
Number of values to write = Z2.
NOTE : If Z2 = 0, nothing will be written (it'll be like the code isn't
executed).
Small note :
------------
As the sign of the address increment and the value increment are shared, you
MUST start
from the 1st address when using a positive value increment, and start from
the last address
when using a negative value increment.
------------------------------------------
// Type "z4 - Size 3" : Memory Copy //
------------------------------------------
These codes were 'created' by me (kenobi).
The only way to use them is to enter and enable the 'Enablers' codes.
You also HAVE TO add the Master Code flag to these Enabler codes'
indentifier
(or to include it into the (m) code), else they won't work properly.
Finally, the 'Enabler' codes and the actual codes must be entered
separately.
They should work on ANY AR (at least up to version 1.14b).
A - Memory Copy Without Pointer Support :
-----------------------------------------
Enabler (must be on!) :
04001E48 48000769
040025B0 5525043E
040025B4 4BFFF644
Exemple of byte copy :
00000000 86393FA8
80393FA0 00000001
Here is how it works :
00000000 8XXXXXXX
YYYYYYYY 0000ZZZZ
8XXXXXXX = [Destination address] OR 0x06000000.
YYYYYYYY = [Source address].
ZZZZ = number of bytes to copy (0x0000 will copy 0 byte, 0xFFFF will copy
65535 bytes).
Important : the 16-bits number before ZZZZ MUST BE '0000', else it'll create
errors !!!
So, if you follow what I explained, you can see that my code exemple will
copy 2 bytes,
from 80393FA0 to 80393FA8.
B - Memory Copy With Pointers Support :
---------------------------------------
Enabler (must be on!) :
04001E48 48000769
040025B0 5525043E
040025B4 2C060000
040025B8 4182000C
040025BC 80630000
040025C0 80840000
040025C4 4BFFF634
With this code, if you put any data in the 8 upper bits of the value, the AR
will use
the addresses in the code as pointers addresses
Exemple :
00000000 86002F04
80002F00 01000138
Important : the 8-bits number before ZZZZ MUST BE '00', else it'll create
errors !!!
As the value start with '01' (could have been anything, but '00'), the AR
will load
the 32bits value at 80002F00 and use it as the source address, then load the
32bits
value at 80002F04 and use it as the destination address, and finally will
copy 138 bytes
from the source address to the destination address.
Note that if you put '00' in the start of the value, the code will work just
like
the 'Memory Copy Without Pointer Support' code.
If you need to add an offset to the pointer addresses, you'll have to do
this trick :
copy the source pointer address to 80002F00, the destination pointer address
to 80002F04,
add the offset values to theses pointer addresses (using the 'Add' code
type), and finally
use the 'Memory Copy with Pointers Support' to copy the bytes.
Exemple :
00000000 86002F00 <- Copy the 32bits (=4 bytes) source pointer address
804C8268 00000004 from 804C8268 to 80002F00.
00000000 86002F04 <- Copy the 32bits (=4 bytes) destination pointer address
804C8268 00000004 from 804C8268 to 80002F04.
84002F00 00000098 <- Add the offset 0x98 to the source pointer address at
80002F00.
84002F04 000001D0 <- Add the offset 0x1D0 to the source pointer address at
80002F04.
4A44F0A8 00000030 <- (if the user press R+Z...).
00000000 86002F04 <- Copy 0x138 bytes from the address stored at 80002F00
(=pointer address+0x98)
80002F00 01000138 to the address stored at 80002F04 (=pointer address +
0x1D0).
________________
----------------
| Normal Codes |
----------------
For any "Normal Codes", you have :
SubType = (ADDRESS >> 30) AND 0x03.
Type = (ADDRESS >> 27) AND 0x07.
Size = (ADDRESS >> 25) AND 0x03.
(usually, Size 0 = 8bits, Size 1 = 16 bits, Size 2 = 32 bits.
For some codes, Size 3 = Floating point single precision)
------------
// Type 0 //
------------
--------------------------------------
// SubType 0 : Ram write (and fill) // (can be called "00", "01" and "02")
--------------------------------------
1 line code.
0.0.x
-----
0wXXXXXX Y1Y2Y3Y4
(w < 8!)
Address = ((0x0wXXXXXXX) AND 0x01FFFFFF) OR 0x80000000).
Size = (address >> 25) AND 0x03.
If Size = 0 [00] :
fills area [Address ; Address + Y1Y2Y3] with value Y4.
If Size = 1 [02] :
fills area [Address ; Address + (Y1Y2 << 1)] with value Y3Y4.
If Size = 2 [04] :
writes word Y1Y2Y3Y4 to Address.
Examples :
00023000 00000312
will write byte 0x12 to 80023000, 80023001, 80023002, 80023003.
02023000 00011234
will write halfword 0x1234 to 80023000, 80023002.
05023000 12345678
will write halfword 0x12345678 to 81023000.
-------------------------------
// SubType 1 : Write to pointer (can be called "04", "05" and "06")
-------------------------------
1 line code.
0.1.x
-----
1 line code.
4wXXXXXX Y1Y2Y3Y4
(w < 8!)
Address = ((0x4wXXXXXX) AND 0x01FFFFFF) OR 0x80000000.
Size = (Address >> 25) AND 0x03.
Pointer Address = [Word stored at Address].
This code will make the AR load the word stored at the address provided in
the code,
(also called the "Pointer Address"), and check if it's a valid address (ie.
if it's in
the [80000000~81800000[ range). It it is one, it will add an offset to it,
and it will
write the data provided in the code to this new address.
If Size = 0 [40] :
AR will write the byte Y4 at [Pointer Address + Y1Y2Y3].
If Size = 1 [42] :
AR will write the halfword Y3Y4 at [Pointer Address + (Y1Y2 << 1)].
If Size = 2 [44] :
AR will write the word Y1Y2Y3Y4 at [Pointer Address].
REMOVE THE 'VALID ADDRESS' CHECK, AKA 'POINTER MOD' :
-----------------------------------------------------
This code was 'created' by me (kenobi).
The only way to use it is to enter and enable the 'Enabler' code.
You also HAVE TO add the Master Code flag to these Enabler codes'
indentifier
(or to include it into the (m) code), else they won't work properly.
Finally, the 'Enabler' codes and the actual codes must be entered
separately.
It should work on ANY AR (at least up to version 1.14b).
Enabler (must be on) :
04001FA4 48000014
Once you use this code, the 'Write to Pointer' code will stop checking if
the address you
point to is a valid address.
That means that you can write to virtual memory without a TLB (m) code, but
you have to make
sure that the address the pointer code reads is a valid address (else, it'll
crash).
Exemple (courtesy of donny2112) :
04002F0C 7FC39C9C
42002F0C 00010000
42002F0C 03ED0000
42002F0C 04F70000
42002F0C 05BB0000
The first line will write '7FC39C9C' to 80002F0C.
Then, the other lines will write 0x0000 to 0x7FC39C9C+2*1,
0x7FC39C9C+2*0x3ED, 0x7FC39C9C+2*0x4F7,
and finally 0x7FC39C9C+2*0x5BB.
The advantage of this code, over a TLB (m) code, is that it only needs a 1
lines enabler, it is
compatible with all games and all ARs, and it allows you to use 8/16/32bits
ram write.
The downside is that if you point to an invalid address, the GC will just
crash.
If you're not sure that you'll point to a valid address, you can use this
combinaison of code to check
it manually (in this exemple, I make sure that the address is in the
0x80000000~817F0000 range) :
74XXXXXX 80000000 <- If value > 0x80000000
2CXXXXXX 81800000 <- and If value < 0x81800000
44XXXXXX Y1Y2Y3Y4 <- then execute this pointer code.
XXXXXXXX being the address where the Pointer Address is stored.
-----------------------
// SubType 2 : Add code (can be called "08", "09" and "0A")
-----------------------
1 line code.
0.2.x
-----
1 line code.
8wXXXXXX Y1Y2Y3Y4
(w < 8!)
Address = (0x8wXXXXXX AND 0x81FFFFFF).
Size = (Address >> 25) AND 0x03.
if Size = 0 [80] :
Loads byte stored at [Address], adds Y1Y2Y3Y4 to it, and stores the
resulting byte
(= result AND 0xFF) at [Address].
if Size = 1 [82] :
Loads halfword stored at [Address], adds Y1Y2Y3Y4 to it, and stores the
resulting halfword
(= result AND 0xFFFF) at [Address].
if Size = 2 [84] :
Loads word stored at [Address], adds Y1Y2Y3Y4 to it, and stores the result
at [Address].
if Size = 3 [86] :
Loads floating value stored at [Address], adds Y1Y2Y3Y4 (must be a
floating point single precision value)
to it, and stores the result at [Address].
Change ADD to AND :
------------------
This code was 'created' by me (kenobi).
The only way to use it is to enter and enable the 'Enabler' code.
You also HAVE TO add the Master Code flag to these Enabler codes'
indentifier
(or to include it into the (m) code), else they won't work properly.
Finally, the 'Enabler' codes and the actual codes must be entered
separately.
This change is definitive (until you reboot the Game) :
Enable 8-bits AND :
0400200C 7C002038
Enable 16-bits AND :
0400201C 7C002038
Enable 32-bits AND :
0400202C 7C002038
Enable 8~32bits AND :
00000000 8400200C
7C002038 00030004
Change ADD to OR :
------------------
This code was 'created' by me (kenobi).
The only way to use it is to enter and enable the 'Enabler' code.
You also HAVE TO add the Master Code flag to these Enabler codes'
indentifier
(or to include it into the (m) code), else they won't work properly.
Finally, the 'Enabler' codes and the actual codes must be entered
separately.
This change is definitive (until you reboot the Game) :
Enable 8-bits OR :
0400200C 7C002378
Enable 16-bits OR :
0400201C 7C002378
Enable 32-bits OR :
0400202C 7C002378
Enable 8~32bits OR :
00000000 8400200C
7C002378 00030004
Note : you can't mix 'ADD', 'AND' and 'OR' codes for the same code type
(8/16/32bits).
----------------------------------------------
// SubType 3 : Master Code & Write to CCXXXXXX (can be called "0E" and "0F")
----------------------------------------------
1 line code.
0.3.x
-----
1 line code.
CwXXXXXX Y1Y2Y3Y4
(w < 8!)
Address = ((0x6wXXXXXX) AND 0x01FFFFFF) OR 0x80000000).
Size = (Address >> 25) AND 0x03.
If Size = 2 (0.3.2) : Master Code (C4XXXXXX Y1Y2Y3Y4)
-----------------------------------------------------
Y4 = Master Code Number.
0x00 : executed only once, just before the game bootup.
Only one (m) code can have the '00' number (the others will be
skipped),
and it must be the very one in the (m) code list (else it'll be
skipped).
0x01~0x0F : executed continuously during the game execution.
(2 (or more) master codes that have the same Master Code
Number can't
be executed correctly if they are put one just after
another.
Only the first one will be executed, the other(s) will be
skipped).
Y3 = number of codes to execute each time the AR "has the hand".
Y2 AND 0x03 = Master Code Type :
Type 0 : create a branch to SUBROUTINE 1.
(Save : R0 R3 R28 R29 R30 R31)
Type 1 : backup 4 asm lines from the game, and write a Branch to MAIN
ROUTINE.
(Save : R3 R28 R29 R30 R31, Destroys : R0?)
Type 2 : create a branch to 1 copy of SUBROUTINE 1.
(Save : R0 R3 R28 R29 R30 R31)
Type 3 : create a branch to MAIN ROUTINE START (will execute the 4 asm lines
backed up
in Type 1, if any).
(Save : R0 R3 R28 R29 R30 R31)
Note : Putting random numbers in Y1 should change the encryption, thus
"signing" your
code (untested).
Note : Don't use the Type 1 alone with a Master Code Number >0, else the AR
will backup its own
hook, and enter an infinite loop. So put a conditional code type make
that this code isn't
executed more than once.
If (Size = 3) AND ((address AND 0x01FFFFFF ) < 0x01000000) (0.3.3):
-------------------------------------------------------------------
Write halfword to CCXXXXXX (C6XXXXXX Y1Y2Y3Y4)
----------------------------------------------
Address = 0xCCXXXXXX
Stores the halfword Y3Y4 at Address.
Note : Putting random numbers in Y1Y2 should change the encryption, thus
"signing" your
code (untested).
If (Size = 3) AND ((address AND 0x01FFFFFF ) >= 0x01000000) (0.3.3):
--------------------------------------------------------------------
Write word to CDXXXXXX (C7XXXXXX Y1Y2Y3Y4)
------------------------------------------
Address = 0xCDXXXXXX
Stores the word Y1Y2Y3Y4 at Address.
Note : Parasyte informed me that writing to 0xCDXXXXXX doesn't makes any
sense, and he thinks
it might be some kind of AR bug...
**************************************************
* NOTES FOR ALL CONDITIONAL CODES (TYPE 1 TO 7). *
**************************************************
All the Conditional Codes are 1 line code, but you "need" to add another
line to make them work.
Conditional Code are used to trigger the next code(s) when an event happens,
for exemple give the
player 99 lifes when buttons L+R are pushed, or make the life becomes full
when it reaches 50%
of its value...
They all come in 3 "flavors" : 8, 16 and 32 bits. You select it by changing
the size data in the code.
Reminder : Size = (Address >> 25) AND 0x03
For all the Conditional Codes, you first take the value of the IN GAME data,
and compare it to
the value provided in the CODE data. The result, which should be read as
'True' (or 'False'), will
tell if the Conditional Code will activate the next codes.
Anyway, Conditional Codes should be used by advanced code makers.
And don't ask for the "paddle" values, they seem to change for every game...
So find them yourself :-)
The number I give as exemples has been made using BYTE size :
08XXXXXX YYYYYY is the "If equal execute next code" generic value for a BYTE
comparison.
For halfwords, it'll be 0AXXXXXX YYYYYYYY, and for words 0CXXXXXX
YYYYYYYY...
--------------------------
// Type 1 : If equal... // (can be called "10", "11" and "12")
--------------------------
1.y.x
-----
08XXXXXX YYYYYYYY
(w >= 8!)
Subtype 0 [08] : If equal, execute next line (else skip next line).
Subtype 1 [48] : If equal, execute next 2 lines (else skip next 2 lines).
Sybtype 2 [88] : If equal, execute all the codes below this one in the same
row (else execute
none of the codes below).
Subtype 3 [C8] : While NOT EQUAL,turn off all codes (infinite loop on the
code).
------------------------------
// Type 2 : If NOT equal... // (can be called "20", "21" and "22")
------------------------------
2.y.x
-----
10XXXXXX YYYYYYYY
Subtype 0 [10] : If NOT equal, execute next line (else skip next line).
Subtype 1 [50] : If NOT equal, execute next 2 lines (else skip next 2
lines).
Sybtype 2 [90] : If NOT equal, execute all the codes below this one in the
same row (else execute
none of the codes below).
Subtype 3 [D0] : While EQUAL, turn off all codes (infinite loop on the
code).
------------------------------------
// Type 3 : If lower... (signed) // (can be called "30", "31" and "32")
------------------------------------
Signed means :
For Bytes : values go from -128 to +127.
For Halfword : values go from -32768/+32767.
For Words : values go from -2147483648 to 2147483647.
For exemple, for the Byte comparison, 7F (127) will be > to FFFFFFFF (-1).
You HAVE to enter a 32bits signed number as value, even if you just want to
make an halfword
comparison. That's because 0000FFFF = 65535, and FFFFFFFF = -1).
You could choose any value (for exemple, +65536 for halfword code, but the
result will be always True
(or always False if you choose -65537...).
3.y.x
-----
18XXXXXX YYYYYYYY
* WARNING * WARNING * WARNING * WARNING * WARNING * WARNING * WARNING *
WARNING * WARNING *
If you used a "byte" size, this Type 3 code will actually be a "If lower...
(UNSIGNED)" !
That means, no signed comparison for byte values !!! (AR bug?)
* WARNING * WARNING * WARNING * WARNING * WARNING * WARNING * WARNING *
WARNING * WARNING *
Subtype 0 [18] : If lower, execute next line (else skip next line).
Subtype 1 [58] : If lower, execute next 2 lines (else skip next 2 lines).
Sybtype 2 [98] : If lower, execute all the codes below this one in the same
row (else execute
none of the codes below).
Subtype 3 [D8] : While higher, turn off all codes (infinite loop on the
code).
Note 1 : For 8 and 16 bits codes, you *could* fill the unused numbers in the
Value to change
the encrypted code, and "sign" them (unverified).
------------------------------------
// Type 4 : If higher... (signed) // (can be called "40", "41" and "42")
------------------------------------
Signed means :
For Bytes : values go from -128 to +127.
For Halfword : values go from -32768/+32767.
For Words : values go from -2147483648 to 2147483647.
For exemple, for the Byte comparison, 7F (127) will be > to FFFFFFFF (-1).
You HAVE to enter a 32bits signed number as value, even if you just want to
make an halfword
comparison. That's because 0000FFFF = 65535, and FFFFFFFF = -1).
You could choose any value (for exemple, +65536 for halfword code, but the
result will be always True
(or always False if you choose -65537...).
4.y.x
-----
20XXXXXX YYYYYYYY
* WARNING * WARNING * WARNING * WARNING * WARNING * WARNING * WARNING *
WARNING * WARNING *
If you used a "byte" size, this Type 4 code will actually be a "If lower...
(UNSIGNED)" !
That means, no signed comparison for byte values !!! (AR bug?)
* WARNING * WARNING * WARNING * WARNING * WARNING * WARNING * WARNING *
WARNING * WARNING *
Subtype 0 [20] : If higher, execute next line (else skip next line).
Subtype 1 [60] : If higher, execute next 2 lines (else skip next 2 lines).
Sybtype 2 [A0] : If higher, execute all the codes below this one in the same
row (else execute
none of the codes below).
Subtype 3 [E0] : While lower, turn off all codes (infinite loop on the
code).
Note 1 : For 8 and 16 bits codes, you *could* fill the unused numbers in the
Value to change
the encrypted code, and "sign" them (unverified).
-------------------------------------
// Type 5 : If lower... (unsigned) // (can be called "50", "51" and "52")
-------------------------------------
Unsigned means :
For Bytes : values go from 0 to +255.
For Halfword : values go from 0 to +65535.
For Words : values go from 0 to 4294967295.
For exemple, for the Byte comparison, 7F (127) will be < to FF (255).
5.y.x
-----
28XXXXXX YYYYYYYY
Subtype 0 [28] : If lower, execute next line (else skip next line).
Subtype 1 [68] : If lower, execute next 2 lines (else skip next 2 lines).
Sybtype 2 [A8] : If lower, execute all the codes below this one in the same
row (else execute
none of the codes below).
Subtype 3 [E8] : While higher, turn off all codes (infinite loop on the
code).
--------------------------------------
// Type 6 : If higher... (unsigned) // (can be called "60", "61" and "62")
--------------------------------------
Unsigned means :
For Bytes : values go from 0 to +255.
For Halfword : values go from 0 to +65535.
For Words : values go from 0 to 4294967295.
For exemple, for the Byte comparison, 7F (127) will be < to FF (255).
6.y.x
-----
30XXXXXX YYYYYYYY
Subtype 0 [30] : If higher, execute next line (else skip next line).
Subtype 1 [70] : If higher, execute next 2 lines (else skip next 2 lines).
Sybtype 2 [B0] : If higher, execute all the codes below this one in the same
row (else execute
none of the codes below).
Subtype 3 [F0] : While lower, turn off all codes (infinite loop on the
code).
------------------------
// Type 7 : If AND... // (can be called "70", "71" and "72")
------------------------
(if the result of ANDing the IN GAME and IN CODE values is <>0)
7.y.x
-----
38XXXXXX YYYYYYYY
Subtype 0 [38] : If AND, execute next line (else skip next line).
Subtype 1 [78] : If AND, execute next 2 lines (else skip next 2 lines).
Sybtype 2 [B8] : If AND, execute all the codes below this one in the same
row (else execute
none of the codes below).
Subtype 3 [F8] : While NOT AND, turn off all codes (infinite loop on the
code).
* THE END *