MESS OUT: Difference between revisions
mNo edit summary |
|||
(One intermediate revision by one other user not shown) | |||
Line 100: | Line 100: | ||
==Command 9: Write Japanese text== | ==Command 9: Write Japanese text== | ||
Same as command 8 but for Japanese characters. | Same as command 8 but for Japanese characters. | ||
Upper-case ASCII (0x20-0x5F) and characters 0x80-0xFE are handled identically to command 8.<br /> | |||
Characters 0x00-0x19 are hiragana with diacritic marks.<br /> | |||
Characters 0x60-0x7A are katakana with diacritic marks.<br /> | |||
The tiles used for this are all taken from page 0, irrespective of the selected fontset.<br /> | |||
0x00-0x1F map to 0x080-0x09F (upper) and 0x0C0-0x0DF (lower).<br /> | |||
0x60-0x7F map to 0x0A0-0x0BF (upper) and 0x0E0-0x0FF (lower).<br /> | |||
==Command A: Call sub list== | ==Command A: Call sub list== | ||
Line 108: | Line 116: | ||
dc.l SUBLIST | dc.l SUBLIST | ||
</syntaxhighlight> | </syntaxhighlight> | ||
'''The call stack seems to be 5 addresses deep.''' Be sure to avoid nesting too many list calls. | |||
==Command B: Return from sub list== | ==Command B: Return from sub list== |
Latest revision as of 02:03, 3 March 2021
MESS_OUT ($C004CE): Generic VRAM output
MESS_OUT executes the command buffer in the System ROM's work RAM that starts at BIOS_MESS_BUFFER (constant $10FF00) and goes up to BIOS_MESS_POINT (longword variable $10FDBE). This can be used to display text messages, boxes... on the fix layer (is it used for sprites too ?).
MESS_OUT is automatically called in the SYSTEM_INT1 call. BIOS_MESS_BUSY ($10FDC2.b) can be used to prevent this (when setting up the command list, for example).
The command buffer can be made of commands directly, or pointers to lists in ROM or user RAM. To put commands directly into the buffer, a null pointer ($00.l) needs to be placed before the command list.
Commands are always words.
Commands
Command 0: End of command list
1 word: $0000
Command 1: Set data format
2 words:
Bit | 15 | 14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 |
Def | 0 | Byte/Word | End code/Data size | $01 |
This command specifies if the data will be in bytes or words, and if the size is defined or if an end code has to be reached.
- If data is in bytes (bit 9 = 0)
- The next word's upper byte will be the data's constant upper byte (since VRAM can only be written in words).
- If the data is limited by an end code (bit 8 = 0), the next word's lower byte will be the end code's value.
- If the data size is defined (bit 8 = 1), the next word's lower byte will be the size (so max length is $FF).
- If data is in words (bit 9 = 1)
- If the data is limited by an end code (bit 8 = 0), the next word will be the end code's value.
- If the data size is defined (bit 8 = 1), the next word will be the size (so max length is $FFFF).
Examples:
dc.w $0001,$15FF ;Data will be read in bytes, upper byte will always be $15, end code is $FF.
dc.w $0101,$1520 ;Data will be read in bytes, upper byte will always be $15, data length is $20.
dc.w $0201,$8000 ;Data will be read in words, end code is $8000.
dc.w $0301,$2044 ;Data will be read in words, data length is $2044.
The format stays the same until command 1 is used again.
Command 2: Set auto-increment
- 1 word
Bit | 15 | 14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 |
Def | Auto-inc value | $02 |
Sets the auto-increment value (written to VRAM_MOD ($3C0004)). Is the value sign-extended or not ?
dc.w $2002 ;Set auto-inc to $20
Command 3: Set VRAM address
- 2 words
Sets the VRAM address written to VRAM_ADDR ($3C0000)) to the next word's value.
dc.w $0003,$7201 ;Set VRAM address to $7201
Command 4: Set data address
- 3 words
Sets the output data address to the next longword. Actual writes to VRAM are done with this command.
dc.w $0004
dc.l MESSAGE
Command 5: Add to current VRAM address
- 2 words
Add the next word to the current VRAM address. (This is different from command 2). Useful to skip lines or spaces.
dc.w $0005,$001F ;Add $1F to the current VRAM address
Command 6: Resume data output
- 1 word
Resume data output (instead of using command 4 again, keep going from the last address).
Command 7: Directly define output data
- at least 2 words
Directly define the data (instead of using command 4 to point to it). Don't forget the padding byte if the data ends on an odd address.
dc.w $0007
dc.b $30,$31,$34,$45,$FF
dc.b $00 ;Pad byte
Command 8: Write text with 8x16 font
- at least 2 words
Bit | 15 | 14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 |
Def | Upper data byte (fontset number) | $08 |
Writes with the 8*16 pixels fix font. The commands upper byte will be the data's constant upper byte. The next bytes are data, and the end code is $FF.
The auto-inc is automaticaly set to $20.
dc.w $0108 ;Tileset 1
dc.b $44,$51 ;Data
dc.b $FF ;End code
dc.b $00 ;Pad byte
Command 9: Write Japanese text
Same as command 8 but for Japanese characters.
Upper-case ASCII (0x20-0x5F) and characters 0x80-0xFE are handled identically to command 8.
Characters 0x00-0x19 are hiragana with diacritic marks.
Characters 0x60-0x7A are katakana with diacritic marks.
The tiles used for this are all taken from page 0, irrespective of the selected fontset.
0x00-0x1F map to 0x080-0x09F (upper) and 0x0C0-0x0DF (lower).
0x60-0x7F map to 0x0A0-0x0BF (upper) and 0x0E0-0x0FF (lower).
Command A: Call sub list
- 3 words
Call sub command list. The next longword is a pointer to another command list. Return has to be made using command B.
dc.w $000A
dc.l SUBLIST
The call stack seems to be 5 addresses deep. Be sure to avoid nesting too many list calls.
Command B: Return from sub list
- 1 word
Return to calling command list.
Command C: Repeated data output
- 2 words
Repeat output. The upper byte of the command is the number of times to output the next word.
dc.w $090C,$0147 ;Outputs $0147 9 times
Command D: Repeat and increment output
- 2 words
Repeat and increment output. The upper byte of the command is the number of times to output the next word. The lower byte of the data is incremented each time.
dc.w $040D,$42FE ;Outputs $42FE,$42FF,$4200,$4201
Example code and command lists
(Slightly modified code from SNK's doc, output gives the above screencap.)
bset.b #0,BIOS_MESS_BUSY
movea.l BIOS_MESS_POINT,a0 ;Get current pointer in buffer
move.l #0,(a0)+ ;Direct commands
move.w #3,(a0)+ ;Set VRAM address to $7318
move.w #$7318,(a0)+
move.w #$0301,(a0)+ ;Words, size = 1
move.w #$0001,(a0)+
move.w #$0007,(a0)+ ;Output data
move.w $100000,(a0)+
move.w #0,(a0)+ ;End of this list
move.l #MESSAGE1,(a0)+ ;Pointer to list in ROM
move.l a0,BIOS_MESS_POINT ;Update pointer
bclr.b #0,BIOS_MESS_BUSY ;Ready to go
rts
MESSAGE1:
dc.w $0001 ;Bytes, upper = $00, end code = $FF
dc.w $00FF
dc.w $2002 ;Auto-inc = $20
dc.w $0003 ;VRAM address = $7024
dc.w $7024
dc.w $0004 ;Write "MESSAGE1"
dc.l MS1
dc.w $0005 ;Next line
dc.w $0001
dc.w $0006 ;Resume, write "MESSAGE2"
dc.w $0005 ;Next line
dc.w $0001
dc.w $0007 ;Direct data output
dc.b "MESSAGE3"
dc.w $FF00 ;End code and pad byte
dc.w $0005 ;Next line
dc.w $0001
dc.w $0108 ;8x16 write, fontset 1
dc.b "ABCDE",$FF
dc.w $0005 ;Next line
dc.w $0001
dc.w $0109 ;8x16 write, fontset 1
dc.b 0,1,2,3,4,$FF
dc.w $0005 ;Next 2 lines
dc.w $0002
dc.w $000A ;Sub list calls
dc.l SUB_MESS
dc.w $000A
dc.l SUB_MESS
dc.w $000A
dc.l SUB_MESS
dc.w $0000
MS1:
dc.b "MESSAGE 1",$FF
dc.b "MESSAGE 2",$FF
SUB_MESS:
dc.w $090C ;Outputs $0022 9 times
dc.w $0022
dc.w $070D ;Outputs $0042 7 times with inc
dc.w $0042
dc.w $0005 ;Next line
dc.w $0001
dc.w $000B ;Return