def cli():   __asm__("cli", [ASM_DESCR.CLOBBER_MEMORY])
def sti():   __asm__("sti", [ASM_DESCR.CLOBBER_MEMORY])
def hlt():   __asm__("hlt", [ASM_DESCR.CLOBBER_MEMORY])
def nop():   __asm__("nop", [ASM_DESCR.CLOBBER_MEMORY])
def pause(): __asm__("pause", op=[ASM_DESCR.CLOBBER_MEMORY])

def inb(port: uint16) -> uint8:
    value: uint32
    __asm__(f"""mov edx, {asminp(uint32(port), ASM_DESCR.REG_ANY)}
              in al, dx
              movzx $0, al""",
    out = [asmout(value, ASM_DESCR.OUTPUT_REG)],
    op =[ASM_DESCR.CLOBBER_RDX, ASM_DESCR.CLOBBER_RAX])
    return uint8(value)

def outb(port: uint16, value: uint8):
    __asm__(f"""mov edx, {asminp(uint32(port), ASM_DESCR.REG_ANY)}
              mov eax, {asminp(uint32(value), ASM_DESCR.REG_ANY)}
              out dx, al""",
    op = [ASM_DESCR.CLOBBER_RDX, ASM_DESCR.CLOBBER_RAX])


def inw(port: uint16) -> uint16:
    value: uint32
    __asm__(f"""mov edx, {asminp(uint32(port), ASM_DESCR.REG_ANY)}
              in ax, dx
              movzx eax, ax
              mov {asmout(value, ASM_DESCR.OUTPUT_REG)}, eax""",
    out = [asmout(value, ASM_DESCR.OUTPUT_REG)],
    op =[ASM_DESCR.CLOBBER_RDX, ASM_DESCR.CLOBBER_RAX])
    return uint16(value)

def outw(port: uint16, value: uint16):
    __asm__(f"""mov edx, {asminp(uint32(port), ASM_DESCR.REG_ANY)}
              mov eax, {asminp(uint32(value), ASM_DESCR.REG_ANY)}
              out dx, ax""",
    op = [ASM_DESCR.CLOBBER_RDX, ASM_DESCR.CLOBBER_RAX])

def inl(port: uint16) -> uint32:
    value: uint32
    __asm__(f"""mov edx, {asminp(uint32(port), ASM_DESCR.REG_ANY)}
              in eax, dx
              mov {asmout(value, ASM_DESCR.OUTPUT_REG)}, eax""",
    out = [asmout(value, ASM_DESCR.OUTPUT_REG)],
    op =[ASM_DESCR.CLOBBER_RDX, ASM_DESCR.CLOBBER_RAX])
    return value

def outl(port: uint16, value: uint32):
    __asm__(f"""mov edx, {asminp(uint32(port), ASM_DESCR.REG_ANY)}
              mov eax, {asminp(value, ASM_DESCR.REG_ANY)}
              out dx, eax""",
    op = [ASM_DESCR.CLOBBER_RDX, ASM_DESCR.CLOBBER_RAX])

def io_wait():
    # ATA 规范的 I/O 延时：**连续 4 次读备选状态寄存器 0x3F6**。它是真实设备端口，每次 in
    # 都是一次真实 I/O 往返，稳定给出 ≥400ns（规范要求命令块寄存器两次写之间至少 400ns）。
    # 原先写的是诊断端口 0x80 —— 该端口背后没有设备，在 QEMU 上几乎不消耗时间，等于**没有
    # 延时**：IDE 命令块寄存器（0x1F2..0x1F7）连写会丢写 ⇒ LBA 寄存器残留上一次的值 ⇒ 读到
    # **别的扇区**的数据。现场（参见 0ee5ec5 与 serial.log）：读根目录扇区 8192 却读回 ELF
    # 文件头 ⇒ 目录项全是垃圾 ⇒ app spawn 失败；或 waitDataReady 超时 ⇒ diskio.read 报错。
    __asm__(f"""mov dx, 0x3F6
              in al, dx
              in al, dx
              in al, dx
              in al, dx""",
    op = [ASM_DESCR.CLOBBER_MEMORY, ASM_DESCR.CLOBBER_DX, ASM_DESCR.CLOBBER_RAX])


inline def pushall():
    __asm__(f"""
        push rax
        push rbx
        push rcx
        push rdx
        push rsi
        push rdi
        push rbp
        push r8
        push r9
        push r10
        push r11
        push r12
        push r13
        push r14
        push r15""",
        [ASM_DESCR.CLOBBER_MEMORY]
    )

inline def popall():
    __asm__(f"""
        pop r15
        pop r14
        pop r13
        pop r12
        pop r11
        pop r10
        pop r9
        pop r8
        pop rbp
        pop rdi
        pop rsi
        pop rdx
        pop rcx
        pop rbx
        pop rax""",
        [ASM_DESCR.CLOBBER_MEMORY]
    )


def cpu_switch_context(old_stack: uint64*, new_stack: uint64):
    __asm__(f"""
        push rax
        push rbx
        push rcx
        push rdx
        push rsi
        push rdi
        push rbp
        push r8
        push r9
        push r10
        push r11
        push r12
        push r13
        push r14
        push r15
        mov [{asminp(old_stack, ASM_DESCR.REG_ANY)}], rsp
        mov rsp, {asminp(new_stack, ASM_DESCR.REG_ANY)}
        pop r15
        pop r14
        pop r13
        pop r12
        pop r11
        pop r10
        pop r9
        pop r8
        pop rbp
        pop rdi
        pop rsi
        pop rdx
        pop rcx
        pop rbx
        pop rax""",
        op=[ASM_DESCR.CLOBBER_MEMORY])

def BSSClean():
    __asm__(
        "lea rdi, [rip + __bss_start]\n"
        "lea rcx, [rip + __bss_end]\n"
        "sub rcx, rdi\n"
        "xor eax, eax\n"
        "cld\n"
        "rep stosb",
        op=[ASM_DESCR.CLOBBER_MEMORY, ASM_DESCR.CLOBBER_RDI, ASM_DESCR.CLOBBER_RCX, ASM_DESCR.CLOBBER_RAX]
    )