Hello world (with proper exit), in 32 and 64 bit implementations.

.gitignore

@@ -0,0 +1,4 @@
+*.o
+*~
+\#*
+.\#*

Makefile

@@ -0,0 +1,12 @@
+
+
+hello32:
+	nasm -f elf hello32.s
+	ld -m elf_i386 -o hello32 hello32.o
+
+hello64:
+	nasm -f elf64 hello64.s
+	ld -o hello64 hello64.o
+
+clean:
+	rm -f hello32 hello64 *.o

hello32.s

@@ -0,0 +1,30 @@
+    ;; Hello World Program #1
+    ;; Compile with: nasm -f elf32 hello.s
+    ;; Link with: ld -m elf_i386 -o hello hello.o
+    ;; Run with: ./hello
+
+    ;; sys/syscall.h
+%define     SYS_write	4
+%define     SYS_exit	1
+
+    ;; unistd.h
+%define STDOUT		1
+
+section .data
+    msg     db  "Hello You Beautiful Human", 0Ah
+
+section .text    
+    global _start
+
+_start:
+    mov     edx, 26d            ; Length of the message
+    mov     ecx, msg            ; Address of the message
+    mov     ebx, STDOUT         ; using STDOUT (see definition above)
+    mov     eax, SYS_write      ; Using WRITE in 32-bit mode?
+    int     80h
+
+    mov     ebx, 0
+    mov     eax, SYS_exit
+    int     80h
+    
+    

hello64.s

@@ -0,0 +1,29 @@
+    ;; Hello World Program #1
+    ;; Compile with: nasm -f elf32 hello.s
+    ;; Link with: ld -m elf_i386 -o hello hello.o
+    ;; Run with: ./hello
+
+    ;; sys/syscall.h
+%define     SYS_write	1
+%define     SYS_exit	60
+
+    ;; unistd.h
+%define STDOUT		1
+
+section .data
+    msg     db  "Hello You Beautiful Human", 0Ah
+
+section .text    
+    global _start
+
+_start:
+    mov     rdx, 26d            ; Length of the message
+    mov     rsi, msg            ; Address of the message
+    mov     rdi, STDOUT         ; using STDOUT (see definition above)
+    mov     rax, SYS_write      ; Using WRITE in 32-bit mode?
+    syscall
+
+    mov     rdi, 0
+    mov     rax, SYS_exit
+    syscall
+    

number.s

@@ -0,0 +1,48 @@
+;; Pure assembly, library-free Linux threading demo
+bits 32
+global _start
+
+;; sys/syscall.h
+%define SYS_write	1
+%define SYS_exit	60
+
+;; unistd.h
+%define STDOUT		1
+
+section .data
+hello:	db `Goodbye\10`
+    
+section .text
+_start:
+	mov eax, 9234d              ; The number we're processing; this is
+                                ; the dividend target for the div operation
+    xor esi, esi                ; fast zero
+.loop:
+    mov edx, 0d                 ; counter
+    mov ebx, 10d                ; divisor
+    div ebx
+    add edx, 48d                ; the remainder, plus the ASCII for '0'
+                                ; the quotient is put into eax
+    push edx                    ; onto the stack
+    inc esi                     ; Increment counter
+    cmp eax, 0d                 ; if EAX is zero, we're done.
+    jz .print
+    jmp .loop
+.print:    
+    cmp esi, 0d                  ; if counter is zero, we're done.
+    jz .exit
+    dec esi                     ; decrement the counter
+    pop ecx                ; copy the address of the stack pointer to edx
+    mov edx, 1                  ; One byte
+	mov ebx, STDOUT             ; To stdout
+	int SYS_write          ; Do it!
+    jmp .print
+
+.exit:
+    mov esi, hello                ; move the contents of the stack pointer into ECX
+    mov ecx, 6                  ; One byte
+	mov ebx, STDOUT             ; To stdout
+	mov eax, SYS_write          ; Do it!
+	mov edi, 0                  ; Success!
+	mov eax, SYS_exit           ; Do it!
+	syscall

something.s

@@ -0,0 +1,25 @@
+;; Pure assembly, library-free Linux threading demo
+bits 32
+global _start
+
+;; sys/syscall.h
+%define SYS_write	1
+%define SYS_exit	60
+
+;; unistd.h
+%define STDOUT		1
+
+section .data
+hello:	db `Goodbye\n`
+    
+section .text
+_start:
+    mov esi, hello
+    mov edx, 8d
+    mov edi, STDOUT
+    mov eax, SYS_write
+    syscall
+
+	mov edi, 0                  ; Success!
+	mov eax, SYS_exit           ; Do it!
+	syscall

threads-x86_64.s

@@ -0,0 +1,108 @@
+;; Pure assembly, library-free Linux threading demo
+bits 64
+global _start
+
+;; sys/syscall.h
+%define SYS_write	1
+%define SYS_mmap	9
+%define SYS_clone	56
+%define SYS_exit	60
+
+;; unistd.h
+%define STDIN		0
+%define STDOUT		1
+%define STDERR		2
+
+;; sched.h
+%define CLONE_VM	0x00000100
+%define CLONE_FS	0x00000200
+%define CLONE_FILES	0x00000400
+%define CLONE_SIGHAND	0x00000800
+%define CLONE_PARENT	0x00008000
+%define CLONE_THREAD	0x00010000
+%define CLONE_IO	0x80000000
+
+;; sys/mman.h
+%define MAP_GROWSDOWN	0x0100
+%define MAP_ANONYMOUS	0x0020
+%define MAP_PRIVATE	0x0002
+%define PROT_READ	0x1
+%define PROT_WRITE	0x2
+%define PROT_EXEC	0x4
+
+%define THREAD_FLAGS \
+ CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_PARENT|CLONE_THREAD|CLONE_IO
+
+%define STACK_SIZE	(4096 * 1024)
+
+%define MAX_LINES	1000000	; number of output lines before exiting
+
+section .data
+count:	dq MAX_LINES
+
+section .text
+_start:
+	; Spawn a few threads
+	mov rdi, threadfn
+	call thread_create
+	mov rdi, threadfn
+	call thread_create
+
+.loop:	call check_count
+	mov rdi, .hello
+	call puts
+	mov rdi, 0
+	jmp .loop
+
+.hello:	db `Hello from \e[93;1mmain\e[0m!\n\0`
+
+;; void threadfn(void)
+threadfn:
+	call check_count
+	mov rdi, .hello
+	call puts
+	jmp threadfn
+.hello:	db `Hello from \e[91;1mthread\e[0m!\n\0`
+
+;; void check_count(void) -- may not return
+check_count:
+	mov rax, -1
+	lock xadd [count], rax
+	jl .exit
+	ret
+.exit	mov rdi, 0
+	mov rax, SYS_exit
+	syscall
+
+;; void puts(char *)
+puts:
+	mov rsi, rdi
+	mov rdx, -1
+.count:	inc rdx
+	cmp byte [rsi + rdx], 0
+	jne .count
+	mov rdi, STDOUT
+	mov rax, SYS_write
+	syscall
+	ret
+
+;; long thread_create(void (*)(void))
+thread_create:
+	push rdi
+	call stack_create
+	lea rsi, [rax + STACK_SIZE - 8]
+	pop qword [rsi]
+	mov rdi, THREAD_FLAGS
+	mov rax, SYS_clone
+	syscall
+	ret
+
+;; void *stack_create(void)
+stack_create:
+	mov rdi, 0
+	mov rsi, STACK_SIZE
+	mov rdx, PROT_WRITE | PROT_READ
+	mov r10, MAP_ANONYMOUS | MAP_PRIVATE | MAP_GROWSDOWN
+	mov rax, SYS_mmap
+	syscall
+	ret