Kernel Syscalls & Bare-Metal Stubs

In the AzLang architecture, standard I/O does not rely on the heavy GNU C Library (glibc or musl). Instead, core operations are mapped directly to Linux kernel system calls using lean assembly stubs.

1. System Call Assembly Implementations

The AzLang SDK provides plain fasm assembly sources for Linux x86_64 (fasm defaults to Intel syntax). These files are pulled in by @link via include — they are never assembled on their own.

write.s — Direct sys_write (Syscall 1)

write:
    mov rax, 1      ; sys_write
    syscall
    ret

On x86_64 Linux ABI:

• rdi: File descriptor (FD fd)
• rsi: Buffer pointer (const str val)
• rdx: Buffer byte count (const int size)
• rax: Syscall number (1 for sys_write)

exit.s — Direct sys_exit (Syscall 60)

exit:
    mov rax, 60     ; sys_exit
    syscall

fopen.s — Direct sys_open (Syscall 2)

fopen:
    mov rax, 2      ; sys_open
    syscall
    ret

2. Advantages of Direct Syscalls

• Zero Startup Penalty: No dynamic symbol relocations or runtime linker (ld-linux.so) overhead.
• Microscopic Footprint: Omits megabytes of standard library code that standard C programs pull in.
• Security & Attack Surface Reduction: Eliminates vulnerabilities inside complex libc string/format-string/locale handlers.