fastest strstr

/*
 * Prototype : char *asm_str_str(const char *s1, const char *s2)
 * Input     : s1 (rdi), s2 (rsi)
 * Output    : char * (rax)
 * Clobber   : rbx, rcx
 */

.globl _asm_str_str
_asm_str_str:
    movq    $-0x1, %rcx
.s_len:
    incq    %rcx
    cmpb    $0x0, (%rsi, %rcx, 0x1)
    jne .s_len
    cmpq    $0x0, %rcx
    je  .s_out
.s_loop:
    cmpq    $0x8, %rcx
    jl  .s_cmp_0x1
.s_cmp_0x8:
    movq    -0x8(%rdi, %rcx, 0x1), %rbx
    cmpq    %rbx, -0x8(%rsi, %rcx, 0x1)
    jne .s_inc
    subq    $0x8, %rcx
        cmpq    $0x7, %rcx
    jg  .s_cmp_0x8
.s_cmp_0x1:
    movb    -0x1(%rdi, %rcx, 0x1), %bl
    cmpb    %bl, -0x1(%rsi, %rcx, 0x1)
    jne .s_inc
    decq    %rcx
    jg  .s_cmp_0x1
    jmp .s_out
.s_inc:
    incq    %rdi
    cmpb    $0x0, -0x1(%rdi)
    jne .s_loop
    xorq    %rdi, %rdi
.s_out:
    movq    %rdi, %rax
    ret