- Fixed a bug to do with how SIB operands were parsed

in the assembler.

- Rewrote more of the SuB Suite to use the new calling
convention.

- Fixed a bug to do with SIB operands in the
emulator's disassembler.

- Made the MMV instruction not require a loop prefix
for copying data.

7 files changed, 146 insertions, 39 deletions

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..a843a26
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,4 @@
+.gitignore
+programs/sub-suite/tmp-stuff/
+*.o
+cisc-0.2
diff --git a/assemble.c b/assemble.c
index 713c297..b51fd5c 100644
--- a/assemble.c
+++ b/assemble.c
@@ -495,6 +495,7 @@ token *get_operands(token *t, operand *op, uint64_t address, uint8_t rs, uint8_t
 	for (; t && i < 2; t = t->next) {
 		reg = (old_i != i) ? 0 : reg;
 		got_value = (old_i != i) ? 0 : got_value;
+		is_sib = (old_i != i) ? 0 : is_sib;
 		if (t->subtype == TOK_IND) {
 			brack_done = (t->id == TOK_REG) ? 2 : 1;
 		}
@@ -512,7 +513,7 @@ token *get_operands(token *t, operand *op, uint64_t address, uint8_t rs, uint8_t
 					} else {
 						break;
 					}
-					is_sib = (!stop_comma && op[i].type && op[i].id == MEM_IND);
+					is_sib = (!stop_comma && op[i].type && (op[i].id == MEM_IND || op[i].id == MEM_RIND));
 					got_value = 1;
 				} else {
 					if (!isvalue) {
@@ -561,6 +562,7 @@ token *get_operands(token *t, operand *op, uint64_t address, uint8_t rs, uint8_t
 							if (!is_sib) {
 								op[i].id = (got_value) ? MEM_ZMR : op[i].id;
 							} else if (op[i].id != MEM_SIB) {
+								reg = (op[i].id == MEM_IND) ? 1 : reg;
 								op[i].type = 1;
 								op[i].id = MEM_SIB;
 								/* An invalid scale value was found.
diff --git a/disasm.c b/disasm.c
index 38770da..9e37646 100644
--- a/disasm.c
+++ b/disasm.c
@@ -253,7 +253,7 @@ static int is_2op(uint8_t opcode) {
 
 static void disasm_ortho(struct sux *cpu, WINDOW *w, uint8_t opcode, uint8_t prefix, uint8_t prefix2, uint64_t *value, char *inst_name, char *postfix, operand *op, uint8_t rs, uint8_t thread) {
 	char opr[2][256];
-	char address[2][17];
+	char address[2][18];
 	char *rind[2] = {"", ""};
 	char *sign[2] = {"", ""};
 	char *ind[2] = {"", ""};
@@ -313,7 +313,13 @@ static void disasm_ortho(struct sux *cpu, WINDOW *w, uint8_t opcode, uint8_t pre
 					}
 				}
 			}
-			sprintf(idx[i], "%s%s%s%s", scale[i], rind[0], (op[i].rind[1] != 0xFF) ? "+" : "", rind[1]);
+
+			if (op[i].rind[1] == 0xFF) {
+				rind[1] = (op[i].id == MEM_SIB) ? rind[0] : rind[1];
+				rind[0] = (op[i].id == MEM_SIB) ? "" : rind[0];
+			}
+
+			sprintf(idx[i], "%s%s%s%s", rind[0], (op[i].rind[1] != 0xFF) ? "+" : "", scale[i], rind[1]);
 			switch (op[i].id) {
 				case MEM_AINDR:
 				case MEM_AIND :
diff --git a/programs/sub-suite/libc.s b/programs/sub-suite/libc.s
index b34d2d3..3de1094 100644
--- a/programs/sub-suite/libc.s
+++ b/programs/sub-suite/libc.s
@@ -411,6 +411,60 @@ free:
 ; Caller preserved registers: D, S, F.
 ; Callee preserved registers: B.
 
+;memcpy:
+;;	inc step		;
+;	phb.q			; Preserve B.
+;	phx.q			; Preserve X.
+;	xor b, b		; Reset B.
+;	mov a, f		; Get the size.
+;	and #7			; Get the size, mod 8.
+;	mov x, f		; Preserve the size.
+;	lsr f, #3		; Divide the size by 8.
+;	beq @rem		; The quotient is zero, so copy the remaining number of bytes.
+;@loop:
+;	mov.q (d+8*b), (s+8*b)	; Copy 8 bytes from the source, into the destination.
+;	inb			; Increment the counter.
+;	cmp b, f		; Is the counter the same as the size?
+;	bne @loop		; No, so keep looping.
+;@rem:
+;	lsl f, #3		; Multiply the quotient by 8.
+;	cmp #0			; Is the remainder zero?
+;	beq @end		; Yes, so we're done.
+;	lea b, 8		; No, so set our offset to 8.
+;	sub b, a		; Subtract 8 by the remainder.
+;	and #0			; Reset A.
+;	dec			; Set the bitmask to -1.
+;	lsl b, #3		; Multiply the offset by 8.
+;	lsr b			; Shift the bitmask right by the offset.
+;	mov.q b, (s+f)		; Get 8 bytes from the source.
+;	xor.q b, (d+f)		; Mask out the destination bytes.
+;	and b			; Get The masked source.
+;	xor.q (d+f), a		; Copy the remaining bytes from the source, into the destination.
+;@end:
+;	mov f, x		; Restore the size.
+;	mov a, d		; Set the return value to the destination pointer.
+;	plx.q			; Restore X.
+;	plb.q			; Restore B.
+;	rts			; End of memcpy.
+
+
+;memcpy:
+;	phb.q			; Preserve B.
+;	phx.q			; Preserve X.
+;	phy.q			; Preserve Y.
+;	mov b, f		; Get the size.
+;	mov x, d		; Get the destination.
+;	mov y, s		; Get the source.
+;	txa			; Set the return value to the destination pointer.
+;@copy:
+;	mmv.r			; Copy the size bytes from the source, into the destination.
+;@end:
+;	ply.q			; Restore Y.
+;	plx.q			; Restore X.
+;	plb.q			; Restore B.
+;	rts			; End of memcpy.
+
+
 memcpy:
 	phb.q			; Preserve B.
 	xor b, b		; Reset B.
@@ -433,6 +487,51 @@ memcpy:
 ; Caller preserved registers: D, S, F.
 ; Callee preserved registers: B.
 
+;memset:
+;;	inc step		;
+;	phb.q			; Preserve B.
+;	phx.q			; Preserve X.
+;	lea b, 8		; Set the loop counter to 8.
+;	psh s			; Get the constant.
+;@const_loop:
+;	lsl s, #8		; Make room for the next byte.
+;	mov s, (sp+1)		; Get the next byte.
+;	deb			; Decrement the loop counter.
+;	bne @const_loop		; The counter is non zero, so keep looping.
+;	ins			; Clean up the frame.
+;@init:
+;	mov a, f		; Get the size.
+;	and #7			; Get the size, mod 8.
+;	mov x, f		; Preserve the size.
+;	lsr f, #3		; Divide the size by 8.
+;	beq @rem		; The quotient is zero, so copy the remaining number of bytes.
+;@loop:
+;	mov.q (d+8*b), s	; Set all 8 destination bytes, to the constant value.
+;	inb			; Increment the counter.
+;	cmp b, f		; Is the counter the same as the size?
+;	bne @loop		; No, so keep looping.
+;@rem:
+;	lsl f, #3		; Multiply the quotient by 8.
+;	cmp #0			; Is the remainder zero?
+;	beq @end		; Yes, so we're done.
+;	lea b, 8		; No, so set our offset to 8.
+;	sub b, a		; Subtract 8 by the remainder.
+;	and #0			; Reset A.
+;	dec			; Set the bitmask to -1.
+;	lsl b, #3		; Multiply the offset by 8.
+;	lsr b			; Shift the bitmask right by the offset.
+;	mov.q b, s		; Get 8 bytes from the constant value.
+;	xor.q b, (d+f)		; Mask out the destination bytes.
+;	and b			; Get The masked source.
+;	xor.q (d+f), a		; Copy the remaining bytes from the source, into the destination.
+;@end:
+;	mov f, x		; Restore the size.
+;	mov a, d		; Set the return value to the destination pointer.
+;	plx.q			; Restore X.
+;	plb.q			; Restore B.
+;	rts			; End of memset.
+
+
 memset:
 	phb.q			; Preserve B.
 	xor b, b		; Reset B.
diff --git a/programs/sub-suite/subasm.s b/programs/sub-suite/subasm.s
index 56d13c2..ba8d2fb 100644
--- a/programs/sub-suite/subasm.s
+++ b/programs/sub-suite/subasm.s
@@ -139,8 +139,10 @@ subasm:
 	mov s, (s)		; Get the command ID.
 	cmp s, #7		; Is the command ID greater than the command count?
 	bcs @end		; Yes, so we're done.
-	lsl s, #1		; No, so multiply the command ID by two.
-	lea.w s, (cmd_srt+s)	; Get the pointer, from the command subroutine table.
+;	lsl s, #1		; No, so multiply the command ID by two.
+;	lea.w s, (cmd_srt+s)	; Get the pointer, from the command subroutine table.
+	lea b, cmd_srt		; No, so get the address of the command subroutine table.
+	mov.w s, (b+2*s)	; Get the pointer, from the command subroutine table.
 	and #0			; Reset A.
 	tab			; Reset B.
 	jsr s			; Run the command's subroutine.
diff --git a/programs/sub-suite/subeditor.s b/programs/sub-suite/subeditor.s
index 83fe179..e44feb3 100644
--- a/programs/sub-suite/subeditor.s
+++ b/programs/sub-suite/subeditor.s
@@ -268,12 +268,11 @@ cmd_cpy:
 	sta scr_trow	; Save it for later.
 	lea d, (buffer)	; Get the screen buffer.
 	mov s, y	; Get the cursor index.
-	lea f, 1	; Set the decrement offset to the left flag.
-	jsr find_end	; Find the end of the line.
-	div #maxcol+1	; Get the ending line.
-	ldb scr_str	; Has the screen been scrolled?
-	beq @start	; No, so don't subtract the screen's starting point from the line number.
-@offset:
+;	jsr find_end	; Find the end of the line.
+;	mov f, a	; Save the end of the line.
+;	div #maxcol+1	; Get the ending line.
+	jsr get_endrow	; Get the ending line.
+;	inc step	;
 	sub scr_str	; Offset the row position, back by the screen's starting point.
 @start:
 	sta scr_row	; Set the row position to the end of the line.
@@ -361,24 +360,21 @@ find_end:
 	phb.q		; Preserve B.
 	phx.q		; Preserve X.
 	phy.q		; Preserve Y.
-	phe.q		; Preserve E.
 	mov a, s	; Set the loop counter.
 	div #maxcol+1	; Get the row position.
 	tay		; Save the row position.
 	mov b, d	; Preserve the string pointer.
 	mov x, s	; Preserve the cursor position.
 	mov d, a	;
-	xor s, s	; Reset S.
 @loop:
-	inc d		; Increment the line.
-	mov e, d	; Save the line.
-	jsr getbit	; Get the linewrap bit.
-	mov d, e	; Get the line.
-	cmp #1		; Is this the end of the line?
-	beq @loop	; No, so keep looping.
-@loop2_1:
-	mov a, d	; Get the ending row position.
-	sub a, f	; Decrement the row position by the offset.
+	xor s, s	; Reset S.
+	iny		; Increment the line.
+	mov d, y	; Get the line.
+	jsr getbit	; Is this the end of the line?
+	bne @loop	; No, so keep looping.
+@loop2_init:
+	tya		; Get the ending row position.
+	dec		; Decrement the row position.
 	mul #maxcol+1	; Get the loop counter.
 	cmp a, x	; Is the new loop counter less than the old loop counter?
 	mcc a, x	; Set the loop counter to the old loop counter if so.
@@ -389,7 +385,6 @@ find_end:
 	cmp (d+a-1), #0	; Did we hit the null terminator?
 	bne @loop2	; No, so keep looping.
 @end:
-	ple.q		; Restore E.
 	ply.q		; Restore Y.
 	plx.q		; Restore X.
 	plb.q		; Restore B.
@@ -445,7 +440,6 @@ printc:
 @update1:
 	lea d, (buffer)	; Get the screen buffer.
 	mov s, y	; Get the cursor index.
-	lea f, 1	; Set the decrement offset to one.
 	jsr get_endrow	; Get the ending line.
 	mov d, a	;
 	mov s, scr_trow	; Get the current row position.
@@ -635,7 +629,6 @@ back:
 	sta scr_trow	;
 	lea d, (buffer)	; Get the screen buffer.
 	mov s, y	; Get the cursor index.
-	lea f, 0	; Set the decrement offset to zero.
 	jsr get_endrow	; Get the ending line.
 	sta scr_row	; Set our row position to the end of the line.
 @find_st:
@@ -664,7 +657,6 @@ back:
 @find_end:
 	lea d, (buffer)	; Get the screen buffer.
 	mov s, y	; Get the cursor index.
-	lea f, 1	; Set the decrement offset to one.
 	jsr get_endrow	; Yes, so Get the ending line.
 	mov d, a	;
 	mov s, scr_trow	; Get the current row position.
@@ -890,9 +882,7 @@ shftln:
 	cmp s, f		; No, but is the cursor position greater than, or equal to the shift length?
 	bcc @end		; No, so we're done.
 @find_end:
-	lea f, 1		; Set the decrement offset to the left flag.
 	jsr find_end		; Find the end of the line.
-	mov f, e		; Get the shift length.
 	sub a, c		; Decrement the old end position, if the left flag is set.
 	tab			; Save the old end position for later.
 	mov a, f		; Get the shift length.
@@ -920,9 +910,13 @@ shftln:
 	mov a, bp		; Get the new end position.
 	div #maxcol+1		; Get the new ending line.
 	mov e, a		; Save the ending line.
+	mov r11, f		; Get the shift length.
 	xor f, f		; Reset F.
 	and c, c		; Is the left flag set?
 	meq f, b		; Save the remainder of the ending line if not.
+;	div #maxcol+1		; Get the line length.
+;	dec			; Decrement the line length.
+;	mov bp, a		; Save the line length.
 	tya			; Get the cursor position.
 	div #maxcol+1		; Get the cursor line.
 	mov bp, a		; Save the cursor line.
@@ -933,16 +927,15 @@ shftln:
 	div #maxcol+1		; Get the old ending line.
 	cmp bp, a		; Is the cursor line greater than, or equal to the old ending line?
 	mcs a, bp		; Get the cursor line if so.
-	inc			; Increment the line.
+	cmp r11, #maxcol+1	; Is the shift length greater than, or equal to the max column count?
+	adc #0			; Increment the starting line if so.
 	and c, c		; Is the left flag set?
 	mne a, e		; Set the ending line to the new ending line if so.
 	tab			; Save the ending line.
 	and c, c		; Is the left flag set?
 	bne @inc		; Yes, so increment the starting line.
 	cmp y, b		; No, but is the starting line less than the ending line?
-	bcs @loop		; No, so don't increment.
-	iny			; Yes, so increment the starting line.
-	bra @getbit		; Get the bit.
+	bcc @inc		; Yes, so increment the starting line.
 @loop:
 	xor s, s		; Reset S.
 	cmp y, b		; Is the current line greater than the ending line?
diff --git a/sux.h b/sux.h
index 44794f8..f2c4227 100644
--- a/sux.h
+++ b/sux.h
@@ -19,7 +19,7 @@
 #define STEP_ADDR 0x110
 #define CURSES_BACKSPACE 0x7F
 
-#define copy64 1
+#define copy64
 
 extern uint8_t kbd_rdy;
 extern uint8_t dbg_print_per_inst;
@@ -338,7 +338,7 @@ static int is_ind(uint8_t type) {
 }
 
 static void *memcopy(void *restrict dst, const void *restrict src, unsigned int n) {
-	#if copy64
+	#ifdef copy64
 	uint64_t *d = dst;
 	const uint64_t *s = src;
 	unsigned int r = n % 8;
@@ -348,7 +348,7 @@ static void *memcopy(void *restrict dst, const void *restrict src, unsigned int
 	const uint8_t *s = src;
 	#endif
 	for (; n; *d++ = *s++, n--);
-	#if copy64
+	#ifdef copy64
     	if (r) {
 		uint64_t mask = (-(uint64_t)1 >> ((8 - r) * 8));
 		*d = (*d & ~mask) | (*s & mask);
@@ -609,6 +609,7 @@ static /*inline*/ uint64_t get_ortho_addr(struct sux *cpu, uint8_t opcode, uint8
 			uint64_t tmp_val = 0;
 			if (is_rind) {
 				for (int j = 0; j < 2 && op[i].rind[j] != 0xFF; j++) {
+					int is_index = ((!j && op[i].rind[1] == 0xFF) || j);
 					uint64_t reg;
 					switch (op[i].rind[j]) {
 						case REG_A  : reg =   cpu->a;	break;
@@ -628,11 +629,11 @@ static /*inline*/ uint64_t get_ortho_addr(struct sux *cpu, uint8_t opcode, uint8
 						case REG_R14: reg = cpu->r14;	break;
 						case REG_R15: reg = cpu->r15;	break;
 					}
+
+					reg *= (is_index && op[i].id == MEM_SIB) ? op[i].scale+1 : 1;
 					tmp_val += reg;
 				}
-				if (op[i].id == MEM_SIB) {
-					tmp_val *= op[i].scale+1;
-				} else if (addr_size != 0xFF) {
+				if (addr_size != 0xFF) {
 					switch (addr_size) {
 						case 0: value[i] = (int8_t )value[i]; break;
 						case 1: value[i] = (int16_t)value[i]; break;
@@ -1361,7 +1362,7 @@ static /*inline*/ void exec_ext_inst(struct sux *cpu, uint8_t opcode, uint8_t pr
 			bit_test(cpu, cpu->a, value, thread);
 			break;
 		case MMV_IMP: /* Memory MoVe. */
-			cpu->b = mem_move(cpu, cpu->b, cpu->x, cpu->y, 0, size, thread);
+			cpu->b = mem_move(cpu, cpu->b, cpu->x, cpu->y, 1, size, thread);
 			break;
 		case SWP_A : /* SWaP lower half, with upper half. */
 			cpu->a = swap(cpu, cpu->a, size, thread);