Linux的工作队列

1，先看看工作队列和tasklet的区别：
    （1） 定时器和tasklet：
    Tasklets resemble kernel timers in 3 ways.
      1)They are always run at interrupt time,
      2)they always run on the same CPU that schedules them,
      3)and they receive an unsigned long argument.   Unlike kernel timers, however, you can’t ask to execute the function at a specific time.
     （2） tasklet和工作队列：
     Workqueues are, superficially, similar to tasklets; they allow kernel code to request that a function be called at some future time. There are, however, some significant differences between the two, including:
        1) Tasklets run in software interrupt context with the result that all tasklet code must be atomic. Instead, workqueue functions run in the context of a special kernel process; as a result, they have more flexibility. In particular, workqueue functions can sleep.
        2)Tasklets always run on the processor from which they were originally submitted. Workqueues work in the same way, by default.
        3)Kernel code can request that the execution of workqueue functions be delayed for an explicit interval.
2，工作队列(work queue)是Linux kernel中将工作推后执行的一种机制。这种机制和BH或Tasklets不同之处在于工作队列是把推后的工作交由一个内核线程去执行，因此工作队列的优势就在于它允许重新调度甚至睡眠。工作队列是2.6内核开始引入的机制，在2.6.20之后，工作队列的数据结构发生了一些变化，因此本文分成两个部分对2.6.20之前和之后的版本分别做介绍。
    （1）2.6.0~2.6.19
数据结构：
struct work_struct {
    unsigned long pending;
    struct list_head entry;
    void (*func)(void *);
    void *data