About linux : How-to-Dynamically-Allocate-Memory-Using-Assembly-and-System-Calls-Under-Linux

Question Detail

I’m looking for some good code examples of dynamic memory allocation using an assembly language under Linux and using system calls, not malloc and friends.

What are some of the simplest but effective ways to do this?

On Intel 386+ computers.

Question Answer

On Linux mmap2 is a sensible system call to use for this at a low level. It takes 6 arguments, so in IA32 you can call it using:

mov eax, 192 ; mmap2
xor ebx, ebx ; addr = NULL
mov ecx, 4096 ; len = 4096
mov edx, $7 ; prot = PROT_READ|PROT_WRITE|PROT_EXEC
mov esi, $22 ; flags = MAP_PRIVATE|MAP_ANONYMOUS
mov edi, -1 ; fd = -1
xor ebp, ebp ; offset = 0 (4096*0)
int $80 ; make call

(See the relevant kernel source for details on the parameter passing)

I built this with NASM and verified it worked using strace, which produced:

mmap2(NULL, 4096, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xf77ae000

……………………………………………………
brk(2). And take a look at ELF.
……………………………………………………
An alternative to brk() is to use the mmap() system call, with MAP_ANONYMOUS | MAP_PRIVATE.
……………………………………………………
Use the brk system call to change the end of your data segment.

Take a look here: http://www.linuxjournal.com/article/6390 to understand what you’re doing.