深刻理解libaio 接口

也許你們都會使用libaio接口，但它和內核是如何交互的呢？內核的機制又是怎樣的呢？下面就一塊兒跟蹤下主要的流程。

登堂入室：系統調用

依賴的頭文件

#include <errno.h>
#include <sys/syscall.h>
#include <unistd.h>

主要的函數：

/* Actual syscalls */
int io_setup(int maxevents, io_context_t *ctxp) {
    return syscall(__NR_io_setup, maxevents, ctxp);
}

int io_destroy(io_context_t ctx) {
    return syscall(__NR_io_destroy, ctx);
}

int io_submit(io_context_t ctx, long nr, struct iocb *ios[]) {
    return syscall(__NR_io_submit, ctx, nr, ios);
}

int io_cancel(io_context_t ctx, struct iocb *iocb, struct io_event *evt) {
    return syscall(__NR_io_cancel, ctx, iocb, evt);
}

(待跟蹤問題：io_get_events() 又是怎樣實現的呢？）

系統調用的實現

那麼上面的系統調用號又具體對應到什麼代碼呢？讓咱們先看看系統調用相關的背景知識：

http://lkml.kernel.org/r/<20110529191055.GC9835%40elte.hu>;

The x86 architecture has quite a few different ways to jump into
kernel code. Most of these entry points are registered in
arch/x86/kernel/traps.c and implemented in arch/x86/entry/entry_64.S
for 64-bit, arch/x86/entry/entry_32.S for 32-bit and finally
arch/x86/entry/entry_64_compat.S which implements the 32-bit compatibility
syscall entry points and thus provides for 32-bit processes the
ability to execute syscalls when running on 64-bit kernels.

The IDT vector assignments are listed in arch/x86/include/asm/irq_vectors.h.

Some of these entries are:

• system_call: syscall instruction from 64-bit code.

• entry_INT80_compat: int 0x80 from 32-bit or 64-bit code; compat syscall
  either way.

• entry_INT80_compat, ia32_sysenter: syscall and sysenter from 32-bit
  code

• interrupt: An array of entries. Every IDT vector that doesn't
  explicitly point somewhere else gets set to the corresponding
  value in interrupts. These point to a whole array of
  magically-generated functions that make their way to do_IRQ with
  the interrupt number as a parameter.

• APIC interrupts: Various special-purpose interrupts for things
  like TLB shootdown.

• Architecturally-defined exceptions like divide_error.

  接着看看異步IO系統調用涉及到的一些系統調用編號,參考arch/ia64/include/uapi/asm/unistd.h文件：

  define __NR_io_setup                   1238
  #define __NR_io_destroy                 1239
  #define __NR_io_getevents               1240
  #define __NR_io_submit                  1241
  #define __NR_io_cancel                  1242
  #define __NR_epoll_create               1243
  #define __NR_epoll_ctl                  1244
  #define __NR_epoll_wait                 1245

那麼這個系統調用編號又是怎樣和功能代碼相管理的呢？

通用入口

ENTRY(entry_SYSCALL_64)
        UNWIND_HINT_EMPTY
        /*
         * Interrupts are off on entry.
         * We do not frame this tiny irq-off block with TRACE_IRQS_OFF/ON,
         * it is too small to ever cause noticeable irq latency.
         */

        swapgs
        /*
         * This path is only taken when PAGE_TABLE_ISOLATION is disabled so it
         * is not required to switch CR3.
         */
        movq    %rsp, PER_CPU_VAR(rsp_scratch)
        movq    PER_CPU_VAR(cpu_current_top_of_stack), %rsp

        /* Construct struct pt_regs on stack */
        pushq   $__USER_DS                      /* pt_regs->ss */
        pushq   PER_CPU_VAR(rsp_scratch)        /* pt_regs->sp */
        pushq   %r11                            /* pt_regs->flags */
        pushq   $__USER_CS                      /* pt_regs->cs */
        pushq   %rcx                            /* pt_regs->ip */
GLOBAL(entry_SYSCALL_64_after_hwframe)
        pushq   %rax                            /* pt_regs->orig_ax */

        PUSH_AND_CLEAR_REGS rax=$-ENOSYS

        TRACE_IRQS_OFF

        /* IRQs are off. */
        movq    %rax, %rdi
        movq    %rsp, %rsi
        call    do_syscall_64           /* returns with IRQs disabled */

        TRACE_IRQS_IRETQ                /* we're about to change IF */

基於系統調用表找處處理函數

#ifdef CONFIG_X86_64
__visible void do_syscall_64(unsigned long nr, struct pt_regs *regs)
{
        struct thread_info *ti;

        enter_from_user_mode();
        local_irq_enable();
        ti = current_thread_info();
        if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY)
                nr = syscall_trace_enter(regs);

        /*
         * NB: Native and x32 syscalls are dispatched from the same
         * table.  The only functional difference is the x32 bit in
         * regs->orig_ax, which changes the behavior of some syscalls.
         */
        nr &= __SYSCALL_MASK;
        if (likely(nr < NR_syscalls)) {
                nr = array_index_nospec(nr, NR_syscalls);
                regs->ax = sys_call_table[nr](regs);
        }

        syscall_return_slowpath(regs);
}
#endif

sys_call_table 在哪？在include/uapi/asm-generic/unistd.h

#define __NR_io_setup 0
__SC_COMP(__NR_io_setup, sys_io_setup, compat_sys_io_setup)
#define __NR_io_destroy 1
__SYSCALL(__NR_io_destroy, sys_io_destroy)
#define __NR_io_submit 2
__SC_COMP(__NR_io_submit, sys_io_submit, compat_sys_io_submit)
#define __NR_io_cancel 3
__SYSCALL(__NR_io_cancel, sys_io_cancel)
#define __NR_io_getevents 4
__SC_COMP(__NR_io_getevents, sys_io_getevents, compat_sys_io_getevents)

調用真正幹活的函數:fs/aio.c: sys_io_submit

/* sys_io_submit:
 *      Queue the nr iocbs pointed to by iocbpp for processing.  Returns
 *      the number of iocbs queued.  May return -EINVAL if the aio_context
 *      specified by ctx_id is invalid, if nr is < 0, if the iocb at
 *      *iocbpp[0] is not properly initialized, if the operation specified
 *      is invalid for the file descriptor in the iocb.  May fail with
 *      -EFAULT if any of the data structures point to invalid data.  May
 *      fail with -EBADF if the file descriptor specified in the first
 *      iocb is invalid.  May fail with -EAGAIN if insufficient resources
 *      are available to queue any iocbs.  Will return 0 if nr is 0.  Will
 *      fail with -ENOSYS if not implemented.
 */
SYSCALL_DEFINE3(io_submit, aio_context_t, ctx_id, long, nr,
                struct iocb __user * __user *, iocbpp)
{
        struct kioctx *ctx;
        long ret = 0;
        int i = 0;
        struct blk_plug plug;

        if (unlikely(nr < 0))
                return -EINVAL;

        ctx = lookup_ioctx(ctx_id);
        if (unlikely(!ctx)) {
                pr_debug("EINVAL: invalid context id\n");
                return -EINVAL;
        }

        if (nr > ctx->nr_events)
                nr = ctx->nr_events;

        blk_start_plug(&plug);
        for (i = 0; i < nr; i++) {
                struct iocb __user *user_iocb;

                if (unlikely(get_user(user_iocb, iocbpp + i))) {
                        ret = -EFAULT;
                        break;
                }

                ret = io_submit_one(ctx, user_iocb, false);
                if (ret)
                        break;
        }

參考文檔：

Documentation/process/adding-syscalls.rst