FDS BIOS

From NESdev Wiki
Jump to navigationJump to search

The Famicom Disk System contains a fixed 8KB BIOS at $E000-FFFF. It controls the Famicom at power-on and reset, dispatches the NMI and IRQ, and offers an API for accessing the data on disk.

Power-on and reset

The BIOS contains a reset handler which initialises the Famicom Disk System's registers and checks for an inserted disk. A basic interface is set up to display error messages if disk loading fails. Once the disk has loaded (or the reset flag and type are set to appropriate values), execution jumps to the game's reset vector.

On power-on and reset, the mirroring is set to horizontal, the disk drive motor is turned off, the stack pointer is set to $FF, and the I flag is cleared. System RAM is filled with values used by the BIOS, and PRG RAM is uninitialized, except for parts of it which have files loaded in.

Approval check

The FDS has a trademark security system similar to what Sega used on some of its consoles:

  • The 224-byte string at PPU $2800-$28DF must match the data in the BIOS, starting at $ED37.
  • This data contains an English licensing message stored in the SMB1/Zelda character encoding. The BIOS ensures that it is present in the nametable and displays it on the screen for several seconds after the boot files are loaded.
  • Traditionally, the first file on a disk is a nametable type file loaded into $2800, which is named KYODAKU- (きょだく or 許諾 means approval). It must be one of the boot files, or the license message test will fail (error $20) before it proceeds to run the program.

The license screen test and display can be bypassed by using a boot file to enable NMIs, which will interrupt the boot loading process early.

Zero-page variables

The FDS BIOS uses the zero-page to store temporary values, controller reads, and the states of several write-only registers:

$00..$0F:	Used as temporary variables.
$F1..$F8:	Used by controller reading routines.
[$F9]:		Mirror of $4026.   $FF on reset.
[$FA]:		Mirror of $4025.   $2E on reset.
[$FB]:		Mirror of $4016.   $00 on reset.
[$FC]:		Mirror of $2005/2. $00 on reset.
[$FD]:		Mirror of $2005/1. $00 on reset.
[$FE]:		Mirror of $2001.   $06 on reset.
[$FF]:		Mirror of $2000.   $10 on reset.

Interrupt/Reset vector controls

The FDS BIOS uses 4 bytes at the lower end of the stack page to control the behaviour of interrupt/reset vectors:

[$0100]:	PC action on NMI. $C0 (NMI #3) on reset.
[$0101]:	PC action on IRQ. $80 (BIOS acknowledge and delay) on reset.
[$0102]:	RESET flag. $35 on reset after the boot files have loaded.
[$0103]:	RESET type. $AC = first boot of the game, $53 = the game was soft-reset by the user.

The following vectors are used depending on the controls:

NMI:
 $E19D :	BIOS disable NMI           (if [$0100] = %00xxxxxx)
($DFF6):	Disk game NMI vector #1    (if [$0100] = %01xxxxxx)
($DFF8):	Disk game NMI vector #2    (if [$0100] = %10xxxxxx)
($DFFA):	Disk game NMI vector #3    (if [$0100] = %11xxxxxx)
RESET:
 $EE24 :	BIOS RESET                 (always runs)
($DFFC):	Disk game RESET vector     (if [$0102] = $35 and [$0103] = $53 or $AC)
IRQ:
 $E1D9 :	BIOS disk skip bytes       (if [$0101] = %00xxxxxx)
 $E1CE :	BIOS disk transfer         (if [$0101] = %01xxxxxx)
 $E1EF :	BIOS acknowledge and delay (if [$0101] = %10xxxxxx)
($DFFE):	Disk game IRQ vector       (if [$0101] = %11xxxxxx)

BIOS calls

Disk access routines

  • These routines take one or two direct pointers as arguments. They must be placed directly after the JSR instruction: the return address in the stack is used to fetch the pointers and is subsequently fixed.
  • Zero-page memory at $00-$0F will be affected by these routines.
  • Unlike the vast majority of disk drives, the FDS lacks any kind of intelligent tracking system. All BIOS load and save functions will do access to the whole disk, no matter which data they load/save. A simple way to overcome this problem is to have a custom loading routine, similar to the BIOS one but forcing the # of files to a smaller number than it actually is. That way the later files are not accessed at all and the earlier files load faster. Of course the maximal time is still taken when loading files that are late on the disk.
  • Dummy loads are used (i.e. file contents are ignored) if the destination address is below $2000, excluding the $x200 page. Loads to $0200-$02FF and its mirror $1200-$12FF are permitted, likely to allow a disk to prepare OAM DMA.
  • All non-disk IRQ sources (timer, DMC and APU frame) should be properly disabled before calling any of these routines. The value at [$0101] however, is preserved on entry, and restored on exit.
  • On return of those routines, A = $00 means no error occurred, other number is error #. Main program should test if an error occurred with the BEQ or BNE instruction, BEQ will branch if no error, and BNE will branch if there is an error.
  • The structures defined below are used to identify files & disks in the access routines. The argument pointers should point to these structures in the program.
Address Name Input parameters Output parameters Description
$E1F8 LoadFiles Direct pointer = Disk ID, Direct pointer = File List A = error #, Y = # of files loaded Load files specified by Disk ID into memory from disk. Load addresses are decided by the file's header. Loads to memory locations below $2000 are ignored but the $x200-$x2FF page is permitted.
$E237 AppendFile Direct pointer = Disk ID, Direct pointer = File Header A = error # Append the file data given by Disk ID to the disk. This means that the file is tacked onto the end of the disk, and the disk file count is incremented. The file is then read back to verify the write. If an error occurs during verification, the disk's file count is decremented (logically hiding the written file).
$E239 WriteFile Direct pointer = Disk ID, Direct pointer = File Header, A = file # A = error # Same as "Append File", but instead of writing the file to the end of the disk, A specifies the sequential position on the disk to write the file (0 is the first). This also has the effect of setting the disk's file count to the A value, therefore logically hiding any other files that may reside after the written one.
$E2B7 CheckFileCount Direct pointer = Disk ID, A = # to set file count to A = error # Read in disk's file count, compare it to A, then set the disk's file count to A.
$E2BB AdjustFileCount Direct pointer = Disk ID, A = # to reduce current file count by A = error # Read in disk's file count, decrement it by A, then writes the new value back.
$E301 SetFileCount1 Direct pointer = Disk ID, A = file count minus one = # of the last file A = error # Set the file count to A + 1.
$E305 SetFileCount Direct pointer = Disk ID, A = file count A = error # Set the file count to A.
$E32A GetDiskInfo Direct pointer = Disk ID A = error # Fill a given location with the current Disk ID data.

Low-Level Disk access routines

Address Name Input parameters Output parameters Description
$E445 CheckDiskHeader $00-$01 = Pointer to 10 byte string Checks that the '*NINTENDO-HVC*' string is present on the current disk, then compares the Disk ID structure to data pointed to by Ptr($00). To bypass the checking of any byte, $FF can be placed in the equivalent field of the specified data. If the comparison fails, an appropriate error will be generated.
$E484 GetNumFiles $06 = # of files Read the number of files from the file amount block.
$E492 SetNumFiles A = # of files Write the number of files in A to the file amount block.
$E4A0 FileMatchTest $02-$03 = Pointer to file ID list Use a byte string pointed at by Ptr($02) to tell the disk system which files to load. The file ID's number is searched for in the string. If an exact match is found, [$09] is set to 0 and [$0E] is incremented. If no matches are found after 20 bytes, or a -1 entry is encountered, [$09] is set to -1. If the first byte in the string is -1, the boot ID number is used for matching files (any file ID that is not greater than the boot ID qualifies as a match).
$E4DA SkipFiles $06 = # of files to skip Skip over a specified number of files.

Example code to load files

Load:
   JSR LoadFiles
   .dw DiskID
   .dw LoadList
   BNE _Error        ;Check if there is an error
   RTS
_Error:
   JSR PrintError      ;If so print the error number and message to screen (include side/disk changing prompts)
_sideError:
   LDA $4032
   AND #$01
   BEQ _sideError     ;Wait until disk is ejected
_insert:
   LDA $4032
   AND #$01
   BNE _insert      ;Wait until disk is inserted
   JMP Load

DiskID:
   .db $01        ;Licensee code
   .db "NAM "     ;3-letter code of game, plus space for normal disk
   .db $00        ;Version
   .db $01        ;Disk side
   .db $00        ;Disk number
   .db $00        ;Disk type
   .db $00        ;Unknown

LoadList          ;In this example the files with IDs equal to $02, $03 or $04 will be loaded into memory
   .db $02, $03, $04, $FF

Error list

error#	BIOS message	explanation
------	------------	-----------
$00			No error
$01	disk set	($4032.0) Disk not set
$02	battery		($4033.7) Power supply failure
$03			($4032.2) Disk is write protected
$04			Wrong licensee code
$05			Wrong game ASCII name string or game type
$06			Wrong game version
$07	a,b side	Wrong side number
$08	disk no.	Wrong disk number
$09			Wrong disk type
$10			Wrong value for unknown field
$20	disk trouble	Approval check failed
$21	disk trouble	'*NINTENDO-HVC*' string in Disk info block doesn't match
$22	disk trouble	Disk info block expected
$23	disk trouble	File amount block expected
$24	disk trouble	File header block expected
$25	disk trouble	File data block expected
$27	disk trouble	($4030.4) Block failed CRC
$28	disk trouble	($4030.6) File ends prematurely during read
$29	disk trouble	($4030.6) File ends prematurely during write
$30	disk trouble	($4032.1) Disk is full

Disk ID structure

This is a structure which is comprised of bytes $0F..$18 of the disk's info block (right after the '*NINTENDO-HVC*' string). Disk access routines call CheckDiskHeader to compare the structure from the disk against specified data. If any of the bytes fail the comparison, an appropriate error # is generated. Comparisons of immaterial data can be skipped by placing an $FF byte in the appropriate field of the Disk ID structure (for example, when the ROM BIOS boots a disk, it sets all the fields in the DiskID string to $FF except for disk side #, and disk #, which are set to 0 (so these fields have to match 0)).

CheckDiskHeader compares the following 10 bytes between the disk and the specified data, and can return the corresponding error #'s.

offset	size	error#	description
------	----	------	-----------
0	1	$04	licensee code
1	4	$05	game ASCII name string + game type
5	1	$06	game version
6	1	$07	disk side #
7	1	$08	disk #
8	1	$09	disk type
9	1	$10	unknown

When the Disk ID structure is returned by GetDiskInfo, the above 10 bytes are followed by additional disk information.

A	1		# of files on disk (from the file amount block)

(The following block will appear for as many files as indicated on the disk)

B	1		file ID code
C	8		file name (ASCII)

(The following is present after the last file info block. Disk size is equal to the sum of each file's size entry, plus an extra 261 per file.)

x	1		disk size high byte
x+1	1		disk size low  byte
x+2			-

File List structure

This is a list of 1-byte IDs of files to load. All files that matches any ID in the list are loaded. A list of up to 20 IDs is possible at a time, smaller lists should be terminated by a $ff byte (this implies a file ID can never be $ff).

Multiple files are loaded in the order as they exist on the disk, not in the order of the list.

File Header structure

This structure is specified when a file is to be written to the disk. The first 15 bytes of this structure directly specifies the data to use for generating a file header block. This is followed by 3 bytes which concern the file data block to be written to the disk.

offset	size	description
------	----	-----------
00	1	file number
01	1	file ID code
02	8	file name
0A	2	load address
0C	2	file data size
0E	1	file type ($00 : Program; $01 : Character; $02 : Nametable)
0F	2	source address of file data (NOT written to disk)
11	1	source address type ($00 : RAM, $01 : VRAM, NOT written to disk)
12		-

Other BIOS calls

Address Name Input parameters Output parameters Affected RAM/Registers Description
$E149 Delay132 132 CPU cycle delay.
$E153 Delayms Y = delay in ms (approximate) X, Y Delay routine, where the time in CPU cycles is 1790*Y+5.
$E161 DisPFObj A, $FE Disable rendering for both sprites and background.
$E16B EnPFObj A, $FE Enable rendering for both sprites and background.
$E171 DisObj A, $FE Disable sprite rendering.
$E178 EnObj A, $FE Enable sprite rendering.
$E17E DisPF A, $FE Disable background rendering.
$E185 EnPF A, $FE Enable background rendering.
$E1B2 VINTWait $FF Wait until next VBlank NMI fires, and return (for programs which do "everything in main"). NMI vector selection at $0100 is preserved, but further VBlanks are disabled.
$E7BB VRAMStructWrite Direct pointer = VRAM struct to be written A, X, Y, $00, $01, $FF Set VRAM increment to 1 (clear PPUCTRL/$FF bit 2), and transfer a VRAM struct to VRAM. Read below for information on the structure.
$E844 FetchDirectPtr $00, $01 = pointer fetched A, X, Y, $05, $06 Fetch a direct pointer from the stack (the pointer should be placed after the return address of the routine calling this one), save the pointer at ($00) and fix the return address.
$E86A WriteVRAMBuffer A, X, Y, $0301, $0302 Transfer the VRAM buffer at $0302 to VRAM. Read below for information on the structure.
$E8B3 ReadVRAMBuffer X = start address of read buffer, Y = # of bytes to read A, X, Y Read individual bytes from VRAM to the VRAM buffer. (see notes below)
$E8D2 PrepareVRAMString A = High VRAM address, X = Low VRAM address, Y = string length, Direct pointer = data to be written to VRAM A = $FF : no error, A = $01 : string didn't fit in buffer A, X, Y, $00-$06 Copy pointed data into the VRAM buffer.
$E8E1 PrepareVRAMStrings A = High VRAM address, X = Low VRAM address, Direct pointer = data to be written to VRAM A = $FF : no error, A = $01 : data didn't fit in buffer A, X, Y, $00-$06 Copy a 2D string into the VRAM buffer. The first byte of the data determines the width and height of the following string (in tiles): Upper nybble = height, lower nybble = width.
$E94F GetVRAMBufferByte X = starting index of read buffer, Y = # of address to compare (starting at 1), $00, $01 = address to read from carry clear : a previously read byte was returned, carry set : no byte was read, should wait next call to ReadVRAMBuffer A, X, Y This routine was likely planned to be used in order to avoid useless latency on a VRAM reads (see notes below). It compares the VRAM address in ($00) with the Yth (starting at 1) address of the read buffer. If both addresses match, the corresponding data byte is returned exit with c clear. If the addresses are different, the buffer address is overwritten by the address in ($00) and the routine exit with c set.
$E97D Pixel2NamConv $02 = Pixel Y coord, $03 = Pixel X coord $00 = High nametable address, $01 = Low nametable address A Convert pixel screen coordinates to the corresponding nametable address (assumes no scrolling, and points to first nametable at $2000-$23ff).
$E997 Nam2PixelConv $00 = High nametable address, $01 = low nametable address $02 = Pixel Y coord, $03 = Pixel X coord A Convert a nametable address to the corresponding pixel coordinates (assume no scrolling).
$E9B1 Random X = Zero Page address where the random bytes are placed, Y = # of shift register bytes (normally $02) A, X, Y, $00 This is a shift-register based random number generator which normally takes 2 bytes (using more won't affect random sequence). On reset the program is supposed to write some non-zero values here (BIOS uses writes $d0, $d0), and call this routine several times before the data is actually random. Each call of this routine will shift the bytes right.
$E9C8 SpriteDMA A Do sprite DMA from RAM $0200-$02FF.
$E9D3 CounterLogic A, Y = end zero-page address of counters, X = start zero-page address of counters A, X, $00 This decrements several counters in zero-page. The first counter is a decimal counter 9 -> 8 -> 7 -> ... -> 1 -> 0 -> 9 -> ... Counters 1...A are simply decremented and stays at 0. Counters A+1...Y are decremented when the first counter does a 0 -> 9 transition, and stays at 0.
$E9EB ReadPads $F5 = Joypad #1 data, $F6 = Joypad #2 data, $00 = Expansion #1 data, $01 = Expansion #2 data A, X Read hardwired/expansion port joypads. Expansion port reads should be placed/used elsewhere if needed after calling this routine, otherwise they will be clobbered by other BIOS calls.
$EA0D OrPads $F5 = Joypad #1 data, $F6 = Joypad #2 data, $00 = Expansion #1 data, $01 = Expansion #2 data $F5 = Joypad #1 OR Expansion #1, $F6 = Joypad #2 OR Expansion #2 A Combine inputs from hardwired/expansion port joypads. Intended to be called after ReadPads.
$EA1A ReadDownPads $F5 = Joypad #1 up->down transitions, $F6 = Joypad #2 up->down transitions $F7 = Joypad #1 data, $F8 = Joypad #2 data A, X, $00, $01 Read hardwired joypads, and detect up->down button transitions.
$EA1F ReadOrDownPads $F5 = Joypad #1 up->down transitions, $F6 = Joypad #2 up->down transitions $F7 = Joypad #1 data, $F8 = Joypad #2 data A, X, $00, $01 Read both hardwired/expansion port joypads and detect up->down button transitions.
$EA36 ReadDownVerifyPads $F5 = Joypad #1 up->down transitions, $F6 = Joypad #2 up->down transitions $F7 = Joypad #1 data, $F8 = Joypad #2 data A, X, $00, $01 Read hardwired joypads, and detect up->down button transitions. Data is read until two consecutive reads match to work around DMC glitches.
$EA4C ReadOrDownVerifyPads $F5 = Joypad #1 up->down transitions, $f6 = Joypad #2 up->down transitions $f7 = Joypad #1 data, $f8 = Joypad #2 data A, X, $00, $01 Read both hardwired/expansion port joypads and detect up->down button transitions. Data is read until two consecutive reads match to work around DMC glitches.
$EA68 ReadDownExpPads $F1-$F4 = up->down transitions, $F5-$F8 = Joypad data, in the order: Pad1, Pad2, Expansion1, Expansion2 A, X, $00, $01 Read both hardwired famicom and expansion port joypads, but store their data separately instead of ORing them together like the other routines do. This routine is NOT DMC fortified.
$EA84 VRAMFill A = High VRAM Address (aka tile row #), X = Fill value, Y = # of tile rows OR attribute fill data A, X, Y, $00-$02 This routine does two things : If A < $20, it fills pattern table data with the value in X for 16 * Y tiles. If A >= $20, it fills the corresponding nametable with the value in X and attribute table with the value in Y.
$EAD2 MemFill A = fill value, X = first page #, Y = last page # A, X, Y, $00, $01 Fill RAM pages with the value in A.
$EAEA SetScroll A Set the scroll registers according to values in $FC, $FD and $FF. Should typically be called in VBlank after VRAM updates.
$EAFD JumpEngine A = Jump table entry A, X, Y, $00, $01 The instruction calling this is supposed to be followed by a jump table (16-bit pointers little endian, up to 128 pointers). A is the entry # to jump to, return address on stack is used to get jump table entries.
$EB13 ReadKeyboard A = $FF : no error, A = $00 : no keyboard connected $00-$08 = keyboard data A, X, Y, $00-$08, $FB Attempt to read the Family BASIC Keyboard, returning with A=0 if nothing is connected. Keyboard data is stored starting from row 8 in descending order, with each byte containing column 0's data in the lower nybble and column 1's data in the upper nybble. A set bit means that a key was pressed.
$EBAF LoadTileset A = Low VRAM Address & Flags, Y = Hi VRAM Address, X = # of tiles to transfer to/from VRAM, Direct pointer = source/destination address in RAM A, X, Y, $00-$04 Read/write 2BP and 1BP tilesets to/from VRAM. See appendix below regarding the flags.
$EC22 UploadObject $00-$01 = Pointer to object structure A, X, Y, $02-$0C Upload an object to RAM $0200-$02FF. (see object structure below)


VRAM transfer structure

SIZE   CONTENTS
2      VRAM Address (big endian)
1      bit 0-5 length of data ($0 means a length of 64)
       bit 6 : 0 = copy, 1 = fill
       bit 7 : 0 = increment by 1, 1 = increment by 32
n      Data to copy to VRAM
.....  repeated as many times as needed
1      $FF
  • The main structure is terminated by a $FF byte (High address is always supposed to be in $00..$3F range)
  • $4C is a "call" command. The 2 bytes that follow is the address of a sub-VRAM structure. The sub-structure can call another sub-structure and so on. (Nesting depth is unknown.)
  • $60 is a "return" command. It will terminate a sub-structure.
  • If Fill mode is used, the routine takes only 1 byte of data which is repeated.
  • The interpretation of the length differs from the otherwise similar Stripe Image buffer format.

VRAM buffer notes

The VRAM buffer is located at $0300-$03xx. $300 holds the size of the buffer (maximum), and $301 holds the end index of the buffer. The actual buffer lies at $0302-$03xx, and is of variable length.

  • $0300 is initialized to the value $7D, effectively making the buffer lie at $0300-$037F. It's possible to change the value here to make it bigger or smaller, but the biggest possible value is $FD, making the buffer lie at $0300-$03FF.
  • Format of the buffer is equivalent to the VRAM structure above, except that there are no sub-structures, no increment by 32 flag and no fill flag.
  • For this reason, the VRAM buffer at $0302 can be used as a sub-structure.
  • A call to WriteVRAMBuffer will execute faster than a call to VRAMStructWrite with $0302 as an argument, but both will have the same effect.

Read routines and VRAM buffer

Unlike the write routines which are very complete, the read routines are somewhat incomplete and their functionality have to be completed by the user (which limits their usefulness). Most of what follows is some sort of speculation about the usefulness of the incomplete read routines of the BIOS.

The read buffer is a part of the VRAM buffer at $0300-$03xx, but the same location can't be used to transfer read and writes on the same frame (for example a RMW operation). Instead the user must manually split the buffer in two parts, the "read" buffer and the "write" buffer. The read buffer is probably supposed to always lies after the write buffer, so that when the read buffer is in use, the value in $300 (size of the write buffer) should be adjusted accordingly. The user should manually keep track of how many bytes are used for the read buffer, as well as the starting point of the read buffer (they are argument to the ReadVRAMBuffer function).

The structure of the read buffer itself is trivial - only single bytes are read (there's no runs of data). Very likely the purpose is to read one or a few individual attribute table data, in order to change the color mapping of individual 2x2 squares instead of whole 4x4 square (read-modify-write operation). All reads are mapped to a structure of 3 bytes in the read buffer :

SIZE   CONTENTS
2      VRAM Address (big endian)
1      data

Therfore, for each byte which is read from VRAM, 3 bytes have to be reserved in the read buffer. Once data from VRAM has been read, if it must be written back after a modification, the user need to copy it to the write buffer manually.

The GetVRAMBufferByte was probably designed to prevent reading attribute tables when it has already been read in the past by a call to ReadVRAMBuffer. For example if you don't know if you're modifying another area of the same 4x4 attribute byte as previously (no need to read again), or if you're modifying part of another 4x4 attribute byte (need to read from VRAM). If the routine return with C=0, the accumulator contains the attribute byte that was already read, but if C=1, it means we should first call ReadVRAMBuffer before getting the data we want.

Since reading VRAM for nothing is not a harmful operation, the user can call ReadVRAMBuffer every frame even if nothing has to be read (no need to check flags) which is probably why these routines are less complete than the write buffer related routines.

Load Tileset notes

The flags parameters are as follows:

7  bit  0
---------
AAAA MMIT
|||| ||||
|||| |||+- Fill bit
|||| ||+-- Transfer direction (0 = Write tiles, 1 = Read tiles)
|||| ++--- Bitplane type (see below)
++++------ Low VRAM Address (aka tile # within a row)

        1st bitplane	2nd bitplane     Description
        -----------	-----------      -----------
    0:  data           data+8           Normal 2-bitplane graphics
    1:  data           fill bit         Single bitplane graphics. Fill bit clear : Use colors 0&1  Fill bit set : Use colors 2&3
    2:  fill bit       data             Single bitplane graphics. Fill bit clear : Use colors 0&2  Fill bit set : Use colors 1&3
    3:  data^fill bit  data             Single bitplane graphics. Fill bit clear : Use colors 0&3  Fill bit set : Use colors 1&2

This makes it possible for single bitplane tiles to take all possible color schemes when they end up in VRAM. However, it is not possible to (natively) load single bitplane graphics directly from the disk into VRAM; it should be loaded into PRG-RAM before transferring the data into VRAM. In read mode, all non "data" bitplanes are replaced by dummy reads.

Object structure

SIZE   CONTENTS
1      Object render flag
            $00 - Render as normal
            $01-$7F - Skip rendering
            $80-$FF - Hide object
1      Y position of upper left of object
1      8-bit fractional portion of Y position
1      X position of upper left of object
1      8-bit fractional portion of X position
1      Animation frame of object
2      Object tile arrangement pointer (big endian)
1      Object flags
            Bit 4: 1 to flip object x-wise
            Bit 0: 1 to flip object y-wise
1      Palette of object
1      Upper nybble - height of object, lower nybble - width of object
1      Object index in OAM

X/Y positions directly correspond to X/Y positions in OAM. The fractional position bytes are not used directly by $EC22, so they can actually be used for any purpose. The intended purposes of those bytes were derived from other code in the BIOS.

Object tiles can be arranged in two ways, by specifying a pointer to the arrangement data, or by simply specifying the upper-left tile ID. In both cases, the tiles are arranged column by column.

If the pointer value is less than $100, the value directly specifies the upper left tile ID, which is then added to (animFrame * width * height) to get the actual first tile ID. For example, displaying animation frame $01 of a 3x3 object whose pointer value is $0010 results in the following sprite arrangement:

$19 $1c $1f
$1a $1d $20
$1b $1e $21

Otherwise, the pointer value is an actual pointer to the tile arrangement data. Let's say we're using 2x3 sprites, our pointer is $6000, and we have the following data at $6000:

$6000: $00 $01 $02 $03 $04 $05
$6006: $ff $fe $fd $fc $fb $fa
$600c: $22 $33 $44 $55 $66 $77

For animation frames $00-$02, we get the following tile arrangements:

Frame $00  Frame $01  Frame $02
 $00 $03    $ff $fc    $22 $55
 $01 $04    $fe $fb    $33 $66
 $02 $05    $fd $fa    $44 $77

References