From 85d0949ec7e1c5ce36e6958e7a03fc697853b6ce Mon Sep 17 00:00:00 2001
From: Nero <41307858+nero@users.noreply.github.com>
Date: Mon, 5 Apr 2021 18:59:44 +0000
Subject: [PATCH] uartctrl: Implement CP437 to DEC VT character conversion

---
 Makefile         |   4 +-
 boot/chartab.asm |  66 +++++++++++++++++++++++++
 rom/uartctrl.asm | 125 +++++++++++++++++++++++++++++++++++------------
 3 files changed, 163 insertions(+), 32 deletions(-)
 create mode 100644 boot/chartab.asm

diff --git a/Makefile b/Makefile
index eab91cb..9a0997a 100644
--- a/Makefile
+++ b/Makefile
@@ -83,5 +83,5 @@ qemu-floppy5: fd360.img $(ROMS)
 qemu-hdd: hdd.img $(ROMS)
 	$(QEMU) $(QEMU_ARGS) -boot c -hda hdd.img
 
-qemu-uartctrl: uartctrl.rom
-	$(QEMU) $(QEMU_ARGS) -option-rom uartctrl.rom -serial stdio
+qemu-uartctrl: uartctrl.rom chartab.bs
+	$(QEMU) $(QEMU_ARGS) -option-rom uartctrl.rom -serial stdio -hda chartab.bs
diff --git a/boot/chartab.asm b/boot/chartab.asm
new file mode 100644
index 0000000..d587cca
--- /dev/null
+++ b/boot/chartab.asm
@@ -0,0 +1,66 @@
+	org 0x7C00
+
+init	xor ax, ax
+	mov ss, ax
+	mov sp, ax
+
+	mov ax, 0x0e20
+	xor bx, bx
+	int 0x10
+	mov cx, 0x10
+.hloop	mov al, 0x20
+	int 0x10
+	mov al, cl
+	neg al
+	add al, 0x10
+	call btoa
+	int 0x10
+	loop .hloop
+	mov al, 0x0A
+	int 0x10
+	mov al, 0x0D
+	int 0x10
+
+	mov cx, 0x20
+.line	test cl, 0x0F
+	jnz .loop
+	mov al, cl
+	sar al, 1
+	sar al, 1
+	sar al, 1
+	sar al, 1
+	and al, 0xF
+	call btoa
+	mov ah, 0x0e
+	int 0x10
+	mov al, 0x20
+	int 0x10
+.loop	mov al, cl
+	int 0x10
+	mov al, 0x20
+	int 0x10
+	mov al, cl
+	inc cx
+	and al, 0xF
+	cmp al, 0xF
+	jne .loop
+
+	mov al, 0x0A
+	int 0x10
+	mov al, 0x0D
+	int 0x10
+
+	cmp cx, 0x100
+	jc .line
+
+halt	hlt
+	jmp halt
+
+btoa	add al, 0x30
+	cmp al, 0x3A
+	jc .ret
+	add al, 7
+.ret	ret
+
+	times (0x1FE - ($-$$)) db 0
+	dw 0xAA55
diff --git a/rom/uartctrl.asm b/rom/uartctrl.asm
index 1a8f70d..b5bd775 100644
--- a/rom/uartctrl.asm
+++ b/rom/uartctrl.asm
@@ -7,6 +7,10 @@
 %define clock 115200
 %define baud 9600
 
+%ifndef CODEPAGE
+%define CODEPAGE 437
+%endif
+
 data	equ (uart + 0) ; data register
 ier	equ (uart + 1)
 fcr	equ (uart + 2)
@@ -19,6 +23,7 @@ msr	equ (uart + 6)
 absolute 0xE0 ; data in bios data area
 pstate	resw 1 ; parser resume offset
 parg	resb 1 ; accumulator for numerical CSI arguments
+curcs	resb 1 ; current charset we put remote term in
 section .text
 
 	db 0x55, 0xAA
@@ -72,12 +77,6 @@ init	push ax
 	mov al, 8
 	out dx, al
 
-	; hook boot interrupt
-	xor ax, ax
-	mov ds, ax
-	mov word [4*0x19], int19
-	mov word [4*0x19+2], cs
-
 	; hook video services
 	; keep original ptr in intvec 47h for chaining
 	mov ax, int10
@@ -87,6 +86,9 @@ init	push ax
 	mov word [4*0x47], ax
 	mov word [4*0x47+2], bx
 
+	; reset parser state
+	mov word [0x400+pstate], parse.start
+
 	; hook uart irq
 	mov word [4*(0x08+irq)], intu
 	mov word [4*(0x08+irq)+2], cs
@@ -97,9 +99,6 @@ init	push ax
 	and al, ~(1 << irq)
 	out dx, al
 
-	; reset parser state
-	mov word [0x400+pstate], parse.start
-
 	call greet
 
 	pop ds
@@ -132,6 +131,26 @@ putc	mov ah, al
 	out dx, al
 	ret
 
+	; set charset
+	; IN	AL 	'0', 'A' or 'B'
+setcs	test al, al
+	jz .fret
+	push ds
+	; set DS
+	mov dx, 0x40
+	mov ds, dx
+	cmp al, [curcs]
+	je .ret
+	mov [curcs], al
+	mov al, 0x1B
+	call putc
+	mov al, '('
+	call putc
+	mov al, [curcs]
+	call putc
+.ret	pop ds
+.fret	ret
+
 pdx:	; this double-call is essentially a 4 times repeating loop
 	call .l1
 .l1:	call .l2
@@ -155,29 +174,27 @@ pputc	mov ah, 0x0e
 	int 0x10
 	ret
 
-	; system bootstrap
-int19	sti
-	xor ax, ax
-	int 0x16
-	mov dx, ax
-	call pdx
-	mov al, 0x0A
-	call pputc
-	mov al, 0x0D
-	call pputc
-	jmp int19
-
 	; video services
 int10   cmp ah, 0x0e
 	jne .chain
 
-	push ax
+	push bx
 	push dx
-
+	push ax
+	; activate character set for current char
+	mov bx, csmap
+	cs xlatb
+	call setcs
+	pop ax
+	push ax
+	; look up what we need to write out
+	mov bx, ecmap
+	cs xlatb
 	call putc
 
-	pop dx
 	pop ax
+	pop dx
+	pop bx
 
 .chain  sub sp, 4
 	push bp
@@ -263,7 +280,7 @@ parse	call ugetc
 	je .vt_o
 	jmp parse
 .csi	mov byte [parg], 0 ; reset argument
-.csicon	call ugetc
+.csicon	call ugetc ; number parsing
 	cmp al, 0x30
 	jc parse
 	cmp al, 0x3A
@@ -275,15 +292,17 @@ parse	call ugetc
 	add al, dl
 	mov [parg], al
 	jmp .csicon
-.nonum	cmp al, '~'
+.nonum	cmp al, '~' ; \e[...$c handling
 	jne .letter
-	mov al, [parg]
+	mov al, [parg] ; \e[$n~ lookup $n
 	mov bx, fkeysn
 	jmp .lookup
 .vt_o	call ugetc ; VT100 \eO... sequences
-.letter	sub al, 0x40
+.letter	sub al, 0x40 ; look up special key by letter
 	mov bx, fkeysc
-.lookup	cs xlatb
+.lookup	cmp al, 0x40 ; special key lookup, bx assumed
+	jnc parse
+	cs xlatb
 	cmp al, 0
 	je parse
 	mov dh, al
@@ -293,7 +312,8 @@ parse	call ugetc
 	align 128
 
 	; ascii -> XT scancode map
-	; this assumes US querty layout
+%if CODEPAGE = 437
+	; US Keyboard layout
 kmap	db 0x00,0x1E,0x30,0x2E, 0x20,0x12,0x21,0x22, 0x23,0x0F,0x24,0x25, 0x26,0x1C,0x31,0x18
 	db 0x19,0x10,0x13,0x1F, 0x14,0x16,0x2F,0x11, 0x2D,0x15,0x2C,0x01, 0x2B,0x1B,0x07,0x0C
 
@@ -306,6 +326,51 @@ kmap	db 0x00,0x1E,0x30,0x2E, 0x20,0x12,0x21,0x22, 0x23,0x0F,0x24,0x25, 0x26,0x1C
 	db 0x29,0x1E,0x30,0x2E, 0x20,0x12,0x21,0x22, 0x23,0x17,0x24,0x25, 0x26,0x32,0x31,0x18
 	db 0x19,0x10,0x13,0x1F, 0x14,0x16,0x2F,0x11, 0x2D,0x15,0x2C,0x1A, 0x2B,0x1B,0x29,0x70
 
+	; character map
+	; resolves codepage character to character in charset (see next table)
+ecmap	db 0x00,0x01,0x02,0x03, 0x60,0x05,0x06,0x07, 0x08,0x09,0x0A,0x0B, 0x0C,0x0D,0x0E,0x0F
+	db 0x10,0x11,0x12,0x13, 0x14,0x15,0x16,0x17, 0x18,0x19,0x1A,0x1B, 0x1C,0x1D,0x1E,0x1F
+	db 0x20,0x21,0x22,0x23, 0x24,0x25,0x26,0x27, 0x28,0x29,0x2A,0x2B, 0x2C,0x2D,0x2E,0x2F
+	db 0x30,0x31,0x32,0x33, 0x34,0x35,0x36,0x37, 0x38,0x39,0x3A,0x3B, 0x3C,0x3D,0x3E,0x3F
+
+	db 0x40,0x41,0x42,0x43, 0x44,0x45,0x46,0x47, 0x48,0x49,0x4A,0x4B, 0x4C,0x4D,0x4E,0x4F
+	db 0x50,0x51,0x52,0x53, 0x54,0x55,0x56,0x57, 0x58,0x59,0x5A,0x5B, 0x5C,0x5D,0x5E,0x5F
+	db 0x60,0x61,0x62,0x63, 0x64,0x65,0x66,0x67, 0x68,0x69,0x6A,0x6B, 0x6C,0x6D,0x6E,0x6F
+	db 0x70,0x71,0x72,0x73, 0x74,0x75,0x76,0x77, 0x78,0x79,0x7A,0x7B, 0x7C,0x7D,0x7E,0x7F
+
+	db 0x47,0x7c,0x69,0x62, 0x64,0x60,0x65,0x67, 0x6A,0x6B,0x68,0x6f, 0x6e,0x6c,0x44,0x45
+	db 0x49,0x66,0x46,0x74, 0x76,0x72,0x7b,0x79, 0x7d,0x56,0x5C,0x22, 0x23,0x25,'P' ,'f'
+	db 0x61,0x6D,0x73,0x7a, 0x71,0x51,0x2A,0x3A, 0x3F,0x2C,0x2C,0x3d, 0x3c,0x21,0x2b,0x3b
+	db 0x61,0x61,0x61,0x78, 0x75,0x75,0x75,0x6B, 0x6B,0x75,0x78,0x6B, 0x6a,0x6a,0x6a,0x6b
+
+	db 0x6D,0x76,0x77,0x74, 0x71,0x6E,0x74,0x74, 0x6D,0x6C,0x76,0x77, 0x74,0x71,0x6E,0x76
+	db 0x76,0x77,0x77,0x6D, 0x6D,0x6C,0x6C,0x6E, 0x6E,0x6A,0x6C,0x61, 0x61,0x61,0x61,0x61
+	db 0x61,0x62,0x47,0x70, 0x53,0x73,0x35,0x74, 0x46,0x6A,0x57,0x64, 0x42,0x66,0x65,0x5C
+	db 0x4F,0x67,0x7A,0x79, 0x24,0x25,0x43,0x49, 0x66,0x7E,0x7E,0x21, 0x32,0x32,0x61,0x5F
+
+	; character set map
+	; resolves codepage character to DEC charset (A,B,0,1,<)
+csmap	db 0,0,0,0, '0',0,0,0, 0,0,0,0, 0,0,0,0
+	db 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0
+	db 0,'B','B','B', 'B','B','B','B', 'B','B','B','B', 'B','B','B','B'
+	db 'B','B','B','B', 'B','B','B','B', 'B','B','B','B', 'B','B','B','B'
+
+	db 'B','B','B','B', 'B','B','B','B', 'B','B','B','B', 'B','B','B','B'
+	db 'B','B','B','B', 'B','B','B','B', 'B','B','B','B', 'B','B','B','B'
+	db 'B','B','B','B', 'B','B','B','B', 'B','B','B','B', 'B','B','B','B'
+	db 'B','B','B','B', 'B','B','B','B', 'B','B','B','B', 'B','B','B',0
+
+	db 'A','A','A','A', 'A','A','A','A', 'A','A','A','A', 'A','A','A','A'
+	db 'A','A','A','A', 'A','A','A','A', 'A','A','A','A', 'A','A','B','B'
+	db 'A','A','A','A', 'A','A','A','A', 'A','A','A','A', 'A','A','A','A'
+	db '0','0','0','0', '0','0','0','0', '0','0','0','0', '0','0','0','0'
+
+	db '0','0','0','0', '0','0','0','0', '0','0','0','0', '0','0','0','0'
+	db '0','0','0','0', '0','0','0','0', '0','0','0','0', '0','0','0','0'
+	db '>','>','>','>', '>','>','A','>', '>','>','>','>', '>','>','>','>'
+	db '>','0','0','0', '>','>','>','>', '0','0','0','>', 'A','A','0','0'
+%endif
+
 	; function keys by character
 	; \eO$c, looks up $c to XT scancode
 	; table is like ascii, except it starts at 0x40 '@'