popr rsi
next
+;; LIT is a special word that reads the next "word pointer" and causes it to be
+;; placed on the stack rather than executed.
+LIT:
+ dq .start
+.start:
+ lodsq
+ push rax
+ next
+
+;; BRANCH is the fundamental mechanism for branching. BRANCH reads the next word
+;; as a signed integer literal and jumps by that offset.
+BRANCH:
+ dq .start
+.start:
+ add rsi, [rsi] ; [RSI], which is the next word, contains the offset; we add this to the instruction pointer.
+ next ; Then, we can just continue execution as normal
+
+;; 0BRANCH is like BRANCH, but it jumps only if the top of the stack is zero.
+ZBRANCH:
+ dq .start
+.start:
+ ;; Compare top of stack to see if we should branch
+ pop rax
+ cmp rax, 0
+ jnz .dont_branch
+.do_branch:
+ jmp BRANCH.start
+.dont_branch:
+ add rsi, 8 ; We need to skip over the next word, which contains the offset.
+ next
+
+;; Expects a character on the stack and prints it to standard output.
EMIT:
dq .start
.start:
popr rsi
next
-PUSH_HELLO_CHARS:
+;; Prints a newline to standard output.
+NEWLINE:
+ dq docol
+ dq LIT, $A
+ dq EMIT
+ dq EXIT
+
+;; Read a word from standard input and push it onto the stack as a pointer and a
+;; size. The pointer is valid until the next call to READ_WORD.
+READ_WORD:
dq .start
- .start:
- push $A
- push 'o'
- push 'l'
- push 'l'
- push 'e'
- push 'H'
+.start:
+ mov [.rsi], rsi
+ mov [.rax], rax
+
+.skip_whitespace:
+ ;; Read characters into .char_buffer until one of them is not whitespace.
+ mov rax, 0
+ mov rdi, 0
+ mov rsi, .char_buffer
+ mov rdx, 1
+ syscall
+
+ cmp [.char_buffer], ' '
+ je .skip_whitespace
+ cmp [.char_buffer], $A
+ je .skip_whitespace
+
+.alpha:
+ ;; We got a character that wasn't whitespace. Now read the actual word.
+ mov [.length], 0
+
+.read_alpha:
+ mov al, [.char_buffer]
+ movzx rbx, [.length]
+ mov rsi, .buffer
+ add rsi, rbx
+ mov [rsi], al
+ inc [.length]
+
+ mov rax, 0
+ mov rdi, 0
+ mov rsi, .char_buffer
+ mov rdx, 1
+ syscall
+
+ cmp [.char_buffer], ' '
+ je .end
+ cmp [.char_buffer], $A
+ jne .read_alpha
+
+.end:
+ push .buffer
+ movzx rax, [.length]
+ push rax
+
+ mov rsi, [.rsi]
+ mov rax, [.rax]
+
next
-HELLO:
+;; Takes a string on the stack and replaces it with the decimal number that the
+;; string represents.
+PARSE_NUMBER:
+ dq .start
+.start:
+ pop [.length] ; Length
+ pop rdi ; String pointer
+ mov r8, 0 ; Result
+
+ ;; Add (10^(rcx-1) * parse_char(rdi[length - rcx])) to the accumulated value
+ ;; for each rcx.
+ mov rcx, [.length]
+.loop:
+ ;; First, calcuate 10^(rcx - 1)
+ mov rax, 1
+
+ mov r9, rcx
+ .exp_loop:
+ dec r9
+ jz .break
+ mov rbx, 10
+ mul rbx
+ jmp .exp_loop
+ .break:
+
+ ;; Now, rax = 10^(rcx - 1).
+
+ ;; We need to calulate the value of the character at rdi[length - rcx].
+ mov rbx, rdi
+ add rbx, [.length]
+ sub rbx, rcx
+ movzx rbx, byte [rbx]
+ sub rbx, '0'
+
+ ;; Multiply this value by rax to get (10^(rcx-1) * parse_char(rdi[length - rcx])),
+ ;; then add this to the result.
+ mul rbx
+
+ ;; Add that value to r8
+ add r8, rax
+
+ dec rcx
+ jnz .loop
+
+ push r8
+
+ next
+
+READ_NUMBER:
dq docol
- dq PUSH_HELLO_CHARS
- dq EMIT
- dq EMIT
- dq EMIT
- dq EMIT
- dq EMIT
- dq EMIT
+ dq READ_WORD
+ dq PARSE_NUMBER
dq EXIT
+;; Takes a string (in the form of a pointer and a length on the stack) and
+;; prints it to standard output.
+TELL:
+ dq .start
+.start:
+ mov rbx, rsi
+ mov rcx, rax
+
+ mov rax, 1
+ mov rdi, 1
+ pop rdx ; Length
+ pop rsi ; Buffer
+ syscall
+
+ mov rax, rcx
+ mov rsi, rbx
+ next
+
+;; Exit the program cleanly.
TERMINATE:
dq .start
- .start:
+.start:
mov rax, $3C
mov rdi, 0
syscall
+PUSH_HELLO_CHARS:
+ dq docol
+ dq LIT, $A
+ dq LIT, 'o'
+ dq LIT, 'l'
+ dq LIT, 'l'
+ dq LIT, 'e'
+ dq LIT, 'H'
+ dq EXIT
+
+PUSH_YOU_TYPED:
+ dq .start
+.start:
+ push you_typed_string
+ push you_typed_string.length
+ next
+
+HELLO:
+ dq docol
+ dq LIT, 'H', EMIT
+ dq LIT, 'e', EMIT
+ dq LIT, 'l', EMIT
+ dq LIT, 'l', EMIT
+ dq LIT, 'o', EMIT
+ dq LIT, '!', EMIT
+ dq NEWLINE
+ dq EXIT
+
+;; .U prints the value on the stack as an unsigned integer in hexadecimal.
+DOTU:
+ dq .start
+.start:
+ mov [.length], 0
+ mov [.printed_length], 1
+ pop rax ; RAX = value to print
+ push rsi ; Save value of RSI
+
+ ;; We start by constructing the buffer to print in reverse
+
+.loop:
+ mov rdx, 0
+ mov rbx, $10
+ div rbx ; Put remainer in RDX and quotient in RAX
+
+ ;; Place the appropriate character in the buffer
+ mov rsi, .chars
+ add rsi, rdx
+ mov bl, [rsi]
+ mov rdi, .rbuffer
+ add rdi, [.length]
+ mov [rdi], bl
+ inc [.length]
+
+ ;; .printed_length is the number of characters that we ulitmately want to
+ ;; print. If we have printed a non-zero character, then we should update
+ ;; .printed_length.
+ cmp bl, '0'
+ je .skip_updating_real_length
+ mov rbx, [.length]
+ mov [.printed_length], rbx
+.skip_updating_real_length:
+
+ cmp [.length], 16
+ jle .loop
+
+ ;; Flip buffer around, since it is currently reversed
+ mov rcx, [.printed_length]
+.flip:
+ mov rsi, .rbuffer
+ add rsi, rcx
+ dec rsi
+ mov al, [rsi]
+
+ mov rdi, .buffer
+ add rdi, [.printed_length]
+ sub rdi, rcx
+ mov [rdi], al
+
+ loop .flip
+
+ ;; Print the buffer
+ mov rax, 1
+ mov rdi, 1
+ mov rsi, .buffer
+ mov rdx, [.printed_length]
+ syscall
+
+ ;; Restore RSI and continue execution
+ pop rsi
+ next
+
MAIN:
dq docol
dq HELLO
- dq HELLO
- dq HELLO
+ dq READ_NUMBER, DOTU, NEWLINE
+ dq BRANCH, -8 * 4
dq TERMINATE
segment readable writable
+you_typed_string db 'You typed: '
+.length = $ - you_typed_string
+
+READ_WORD.rsi dq ?
+READ_WORD.rax dq ?
+READ_WORD.max_size = $FF
+READ_WORD.buffer rb READ_WORD.max_size
+READ_WORD.length db ?
+READ_WORD.char_buffer db ?
+
+DOTU.chars db '0123456789ABCDEF'
+DOTU.buffer rq 16 ; 64-bit number has no more than 16 digits in hex
+DOTU.rbuffer rq 16
+DOTU.length dq ?
+DOTU.printed_length dq ?
+
+PARSE_NUMBER.length dq ?
+
;; Return stack
rq $2000
return_stack_top: