IT WORKS

Makefile

@@ -8,6 +8,11 @@ DEBUG_LD_FLAGS := -g
 DEBUG_NASM_FLAGS := -g -F dwarf
 
 # check for avx2 support
+ifeq ($(shell grep -o 'avx2[^ ]*' /proc/cpuinfo | head -n 1),avx512)
+    NASM_FLAGS += -DAVX2
+endif
+
+# check for avx512 support
 ifeq ($(shell grep -o 'avx512[^ ]*' /proc/cpuinfo | head -n 1),avx512)
     NASM_FLAGS += -DAVX512
 endif

src/drawing.asm

@@ -23,17 +23,30 @@ section .rodata
 
 	home_cursor:	db ESC_CHAR, "[H", 0
 
-	statusbar:	db ESC_CHAR, "[30;100m", "Use arrow keys to move cursor, enter to invert cell j/k to change simulation speed, p to       simulation", 0
+	statusbar:	db ESC_CHAR, "[32;100m", "Use arrow keys to move cursor, enter to invert cell j/k to change simulation speed, p to       simulation", 0
 	START_STOP_pos: equ $-statusbar-17
 	
 	
 	start_str:	db "START", 0
 	stop_str:	db "STOP ", 0
 
+	alive_switch_statement:
+		dq step_simulation.die; 0
+		dq step_simulation.die; 1
+		dq step_simulation.live; 2
+		dq step_simulation.live; 3
+		dq step_simulation.die; 4
+		dq step_simulation.die; 5
+		dq step_simulation.die; 6
+		dq step_simulation.die; 7
+		dq step_simulation.die; 8
+
+
 section .text 
 extern print_str
 extern string_copy
 extern memory_set
+extern memory_copy
 extern alloc
 
 global init_gameboard
@@ -99,69 +112,111 @@ print_game_ui:
 	
 	ret
 
-%macro check_if_hashtag 1
-	cmp r8, %1
-	jl +7
-	cmp r9, %1 
-	ja +5
+%macro check_if_hashtag 2
+	cmp %1, r8
+	jl .no_count_%2
+	cmp %1, r9
+	ja .no_count_%2
 	mov r11b, [%1]
 	cmp r11b, '#' 
-	jne +2 
+	jne .no_count_%2
 	inc dl
+	.no_count_%2:
 %endmacro
 
 global step_simulation:
-step_simultion:
+step_simulation:
 	mov rdi, [next_frame_ptr]; destination
 	mov rsi, [gameboard_ptr]; source 
 	mov rcx, [gameboard_size]; number of iterations
-
+	
 	mov r8, rsi; store lowest address posible so we are not checking out of bounds
 	mov r9, rsi
 	add r9, rcx; store higest address posible so we are not checking out of bounds
 
-	mov r10, [term_cols]
+	xor r10, r10
+	mov r10w, [term_cols]
 	;mov r11, [term_rows] this register has been confiscated since i cannot use ah because of error: cannot use high byte register in rex instruction
 
+	sub rcx, r10; remove status bar
+
 	xor rax, rax; this shouldn't be needed but just to be sure
 	xor r11, r11
 	xor rdx, rdx; we will use dl as # counter 
 	.for_every_column_on_gameboard:
-	mov al, [rdi]; NOTE to self if i need extra register i can shift this to ah and free up r11
+	xor dl, dl
+	mov al, [rsi]; NOTE to self if i need extra register i can shift this to ah and free up r11
 	
 	
-	inc rdi
-	check_if_hashtag rdi
-	dec rdi
+	inc rsi
+	check_if_hashtag rsi, 1; check column to the to the right
+	dec rsi
 
-	check_if_hashtag rdi-1
+	dec rsi
+	check_if_hashtag rsi, 2; check the one to the left
+	inc rsi
 
-
-	add rdi, r10
+	add rsi, r10
 	
-	check_if_hashtag rdi
+	check_if_hashtag rsi, 3; check the one to the down
 
+	inc rsi
+	check_if_hashtag rsi, 4; check the one to the down-right
+	dec rsi
+
+	dec rsi
+	check_if_hashtag rsi, 5; check the one to the down-left
+	inc rsi
+
+	sub rsi, r10
+
+
+	sub rsi, r10	
+	check_if_hashtag rsi, 6; check the one to the up
+
+	inc rsi
+	check_if_hashtag rsi, 7; check the one to the up-right
+	dec rsi
+
+	dec rsi
+	check_if_hashtag rsi, 8; check the one to the up-left
+	inc rsi
+
+	add rsi, r10
+
+	cmp al, '#'
+	jne .dead_cell
+
+	jmp [alive_switch_statement+(rdx*8)]
+
+	.die:
+	mov byte [rdi], 0x20; SPACE
+	jmp .end_check
+
+	.live:
+	mov byte [rdi], '#'
+	jmp .end_check
+
+	.dead_cell:
+	cmp dl, 3 
+	jne .fill_space_there
+	mov byte [rdi], '#'
+	jmp .end_check
+	.fill_space_there:
+	mov byte [rdi], 0x20; SPACE
+
+	.end_check:
+	dec rcx
 	inc rdi
-	check_if_hashtag rdi
-	dec rdi
+	inc rsi
+	test rcx, rcx
+	jnz .for_every_column_on_gameboard
 
-	check_if_hashtag rdi-1
+	mov rsi, [next_frame_ptr]; source
+	mov rdi, [gameboard_ptr]; destination 
+	mov rdx, [gameboard_size]; number of iterations
+	sub rdx, r10; remove statusbar
 
-	sub rdi, r10
-
-
-	sub rdi, r10	
-	check_if_hashtag rdi
-
-	inc rdi
-	check_if_hashtag rdi
-	dec rdi
-
-	check_if_hashtag rdi-1
-
-	add rdi, r10
-
-
-	; TODO create jump table
+	call memory_copy
 
 	ret

src/input.asm

@@ -1,8 +1,8 @@
 %include "symbols.asm"
 
 section .bss
-	cursor_rows: RESW 1; TODO DONT FORGET TO INICIALIZE
-	cursor_cols: RESW 1
+	alignb 16
+	termios: RESZ 1; 60 bytes is needed i use 64 for alligment and it is easier to work with
 
 	extern multipurpuse_buf
 
@@ -12,13 +12,16 @@ section .bss
 	extern gameboard_ptr
 
 	extern simulation_running
-
+section .data 
+	cursor_rows: dw 1
+	cursor_cols: dw 1
 section .rodata
 
 	cursor_up: db ESC_CHAR, "[1A", 0
 	cursor_down: db ESC_CHAR, "[1B", 0
 	cursor_right: db ESC_CHAR, "[1C", 0
 	cursor_left: db ESC_CHAR, "[1D", 0
+
 	
 
 	arrow_switch_statement:
@@ -30,73 +33,144 @@ section .rodata
 section .text
 
 extern print_str
+extern step_simulation
+extern unsigned_int_to_ascii
+extern print_game_ui
 
 global handle_user_input
 handle_user_input:; main loop of the program
+	push r12
+
+	lea r12, [multipurpuse_buf]
 
 	.main_loop:
 
+	; put the cursor where it should be 
+	mov rdi, r12; multipurpuse_buf pointer is in r12
+	mov word [rdi], 0x5B1B; will store ESC_CHAR, '[' they have to be in reverse order here due to little endian
+	add rdi, 2
+	push rdi
+	xor rsi, rsi
+	mov si, [cursor_rows]
+	call unsigned_int_to_ascii
+	pop rdi
+	add rdi, rax; add lenght of string to pointer 
+	mov byte [rdi], ';'
+	inc rdi
+	push rdi
+	mov si, [cursor_cols]
+	call unsigned_int_to_ascii
+	pop rdi
+	add rdi, rax
+	mov byte [rdi], 'H'
+	inc rdi
+	mov byte [rdi], 0; null terminate
+
+	mov rdi, r12; multipurpuse_buf pointer is in r12
+	call print_str
+
+	
+
 	xor rax, rax
-	mov qword [multipurpuse_buf], rax; zeroout the buffer
+	mov qword [r12], rax; zeroout the buffer
 
 	mov rax, SYS_POLL
-	mov rdi, STDIN
+	mov dword [r12], STDIN; create pollfd struct
+	mov word [r12+4], POLLIN
+	mov rdi, r12
 	mov rsi, 1; only one file descriptor is provided
-	mov rdx, 0; no timeout. maybe use this for final sleep but run if user inputs something TODO
+	mov rdx, 500; no timeout. maybe use this for final sleep but run if user inputs something TODO
 	syscall
 
 	test rax, rax; SYS_POLL returns 0 when no change happens within timeout
-	jz .no_input 
+	jz .no_input
+
+	xor rax, rax
+	mov qword [r12], rax; zeroout the buffer
 	
 	mov rax, SYS_READ
 	mov rdi, STDIN
-	lea rsi, [multipurpuse_buf]
+	lea rsi, [r12]
 	mov rdx, 8; size of multipurpuse buffer
 	syscall; read user input
 
-	mov rax, [multipurpuse_buf]
-	shr rax, 5; we need only 3 bytes for this inpus sceame
+	cmp rax, EAGAIN
+	je .no_input
 
-	cmp eax, 0x001B5B44; check if input is more than left arrow
-	ja .handle_single_byte_chars
+	mov rax, [r12]
+	
+	cmp eax, 0x00415B1B; check if input is more than left arrow
+	jl .handle_single_byte_chars
+
+	bswap eax
+
+	sub eax, 0x1B5B4100
+	shr eax, 8
+	cmp al, 3
+	ja .no_input
+
+	mov r9w, [term_rows]
+	dec r9w
+	mov r10w, [term_cols]
 
-	sub eax, 0x1B5B41
 	jmp [arrow_switch_statement+(rax*8)]; lets hope this works
 
 	.arrow_up:
 	dec word [cursor_rows]
+	jnz .move_cursor_up
+	inc word [cursor_rows]
+	jmp .no_input
+	.move_cursor_up:
 	lea rdi, [cursor_up]
 	call print_str
 	jmp .no_input
+
 	.arrow_down:
-	inc word [cursor_rows]
+	mov r8w, [cursor_rows]
+	inc r8w
+	cmp word r8w, r9w
+	ja .no_input
+	mov word [cursor_rows], r8w
 	lea rdi, [cursor_down]
 	call print_str
 	jmp .no_input
+
 	.arrow_right:
- 	inc word [cursor_cols]
+ 	mov r8w, [cursor_cols]
+	inc r8w
+	cmp word r8w, r10w
+	ja .no_input
+	mov word [cursor_cols], r8w
 	lea rdi, [cursor_right]
 	call print_str
 	jmp .no_input
+
 	.arrow_left:
 	dec word [cursor_cols]
+	jnz .move_cursor_left 
+	inc word [cursor_cols]
+	jmp .no_input
+	.move_cursor_left:
 	lea rdi, [cursor_left]
 	call print_str
 	jmp .no_input
 
 	.handle_single_byte_chars:
 
-	shr eax, 2; get the char to al 
 
 	cmp al, 0xa; NEWLINE (enter key)
 	jne .check_p
 
 	xor rax, rax; zeroout rax
 	mov ax, [cursor_rows]
-	mul dword [term_cols]
-	add rax, [cursor_cols]
+	dec ax
+	mul word [term_cols]
+	mov cx, [cursor_cols]
+	dec cx
+	add ax, cx
 	
-	lea rdi, [gameboard_ptr+rax]
+	mov rdi, [gameboard_ptr]
+	add rdi, rax
 	mov cl, [rdi]
 	cmp cl, '#'
 	je .hashtag_present
@@ -127,15 +201,92 @@ handle_user_input:; main loop of the program
 
 	.check_k:
 	cmp al, 'k'
-	jne .no_input
+	jne .check_q
 
 	; TODO implement simulation speed
 
+	.check_q:
+	cmp al, 'q'
+	jne .no_input
+	pop r12
+	ret; exit if q pressed
+
 	.no_input:
 
-		
+	mov al, [simulation_running]
+	test al, al
+	jz .dont_step
+	call step_simulation
+	.dont_step:
+	call print_game_ui
 	jmp .main_loop
 
+	pop r12
+	ret
+
+	
+global disable_canonical_mode_and_echo
+disable_canonical_mode_and_echo:
+	
+	mov rax, SYS_IOCTL
+	mov rdi, STDIN
+	mov rsi, TCGETS
+	lea rdx, [termios]
+	syscall 
+
+	; save original termios struct
+	%ifdef AVX2
+		%ifdef AVX512
+			vmovdqa64 zmm0, [termios]
+		%else
+			vmovdqa ymm0, [termios]
+			vmovdqa ymm1, [termios+32]						
+		%endif 
+	%else
+		vmovdqa xmm0, [termios]
+		vmovdqa xmm1, [termios+16]
+		vmovdqa xmm2, [termios+32]
+		vmovdqa xmm3, [termios+64]
+	%endif
+
+	
+	mov eax, [termios+12]; get c_lflag
+	and eax, NOT_ECHO; disable ECHO
+	and eax, NOT_ICANON; disable ICANON
+	mov [termios+12], eax
+
+	mov rax, SYS_IOCTL
+	mov rdi, STDIN
+	mov rsi, TCSETS
+	lea rdx, [termios]
+	syscall 
+
+
+	; load original termios struct
+	%ifdef AVX2
+		%ifdef AVX512
+			vmovdqa64 [termios], zmm0
+		%else
+			vmovdqa [termios], ymm0
+			vmovdqa [termios+32], ymm1						
+		%endif 
+	%else
+		vmovdqa [termios], xmm0
+		vmovdqa [termios+16], xmm1
+		vmovdqa [termios+32], xmm2
+		vmovdqa [termios+64], xmm3
+	%endif
+
+
+	ret
+
+global reset_terminal
+reset_terminal:
+	mov rax, SYS_IOCTL
+	mov rdi, STDIN
+	mov rsi, TCSETS
+	lea rdx, [termios]
+	syscall 
 	ret
 	
 	

src/main.asm

@@ -33,6 +33,11 @@ extern alloc
 extern init_gameboard
 extern print_game_ui
 
+extern handle_user_input
+
+extern disable_canonical_mode_and_echo
+extern reset_terminal
+
 global _start
 _start:
 	; get terminal dimensions
@@ -82,7 +87,11 @@ _start:
 
 	call print_game_ui	
 
+	call disable_canonical_mode_and_echo
 
+	call handle_user_input
+
+	call reset_terminal
 
 
 

src/str.asm

@@ -24,7 +24,7 @@ print_str: ; takes pointer to string in rdi and retuns in rax
     ret
 
 global unsigned_int_to_ascii
-unsigned_int_to_ascii: ; takes pointer to array in rdi and value stored in rsi DOES NOT BOUNDS CHECK
+unsigned_int_to_ascii: ; takes pointer to array in rdi and value stored in rsi DOES NOT BOUNDS CHECK return len of string in rax
     xor r11, r11
     mov rcx, 10
     mov rax, rsi
@@ -154,52 +154,16 @@ memory_copy:; takes  destination in rdi, source in rsi and lenght in rdx
 
 	mov r9, rdi
 
-	cmp rdx, 16
-	jnl .write_16_or_more_bytes 
-	mov rcx, rdx 
-	jmp .write_less_than_16_bytes
-	.write_16_or_more_bytes:
-	mov rax, rdi; move destination to rax
-	and rax, 0xF; offset is stored in rax
+	mov rcx, rdx
+	shr rcx, 3; calculate how many 8 byte chunks we need
+	and rdx, 0x7; calculate the reminder
 
-	
-	test al, al; check if resault is 0
-	jz .addr_is_16_Byte_alligned
-	
+	rep movsq; move as many 8 byte chunks as posible
 
-	mov cl, 16
-	sub cl, al; now offset to first higher 16 byte alligned address is stored in r8
-	movzx rcx, cl; remove ani posible garbage
-	
+	mov rcx, rdx
 
-	.write_less_than_16_bytes:
-	sub rdx, rcx; we will write these bytes now
-	
-	rep movsb
+	rep movsb; move the rest
 
-	.addr_is_16_Byte_alligned:
-	mov r10, rdx
-	shr r10, 4; set it to how many 128bit(16Byte) chunk we need 
-	test r10, r10; check if we need to write aditional 16 bytes at all
-	jz .function_exit
-		
-	.move_16_bytes:
-	movdqa xmm8, [rsi]
-	movdqa [rdi], xmm8
-	add rdi, 16
-	add rsi, 16
-	sub rdx, 16
 
-	cmp rdx, 16; test if rdx is less than 16
-	jge .move_16_bytes
-
-	.function_exit:
-
-	test rdx, rdx; test if rdx is 0
-	jz .true_function_exit
-	movzx rcx, dl
-	jmp .write_less_than_16_bytes
-
-	.true_function_exit:
-	mov rax, r9; return pointer to memory area same as memset in libc
+	mov rax, r9; return pointer to memory area same as memcpy in libc
 	ret	

src/symbols.asm

@@ -10,9 +10,16 @@ STDIN		equ 0
 STDOUT		equ 1
 
 TIOCGWINSZ	equ 0x5413
-POLLIN		equ 1 
+TCGETS		equ 0x5401
+TCSETS		equ 0x5402
 F_SETFL		equ 4 
 O_NONBLOCK	equ 2048
+POLLIN		equ 0x0100; compensate for litle endian
+
+NOT_ECHO	equ -9
+NOT_ICANON	equ -3
+
+EAGAIN		equ -11
 
 ASCII_ZERO 	equ 48
 ESC_CHAR 	equ 27