*看了一段時間Linux內核源代碼了,經常會在代碼中看到down()、up()、spin_lock()、spin_unlock()、read_lock()、write_lock()、read_unlock()、write_unlock()等函數。本篇就先來看down()、up()是幹什麼的。。。它們的底層都是如何實現的。。。→_→*
down()(P操作)
內核中通過信號量(semaphore)來實現進程間對共享資源的互斥訪問,提供了down()函數(P操作)和up()函數(V操作)。
- 內核中信號量的數據結構
//linux-2.4.0\include\asm-i386\Semaphore.h struct semaphore { atomic_t count;//計數器,表示可用資源的數量 int sleepers;//等待進程的數量(其實只代表有沒有進程等待) wait_queue_head_t wait;//進程的等待隊列 #if WAITQUEUE_DEBUG long __magic; #endif };
- 初始化信號量
#if WAITQUEUE_DEBUG # define __SEM_DEBUG_INIT(name) \ , (int)&(name).__magic #else # define __SEM_DEBUG_INIT(name) #endif //初始化count與等待隊列 #define __SEMAPHORE_INITIALIZER(name,count) \ { ATOMIC_INIT(count), 0, __WAIT_QUEUE_HEAD_INITIALIZER((name).wait) \ __SEM_DEBUG_INIT(name) } //初始化信號量 #define __MUTEX_INITIALIZER(name) \ __SEMAPHORE_INITIALIZER(name,1) #define __DECLARE_SEMAPHORE_GENERIC(name,count) \ struct semaphore name = __SEMAPHORE_INITIALIZER(name,count) //聲明初始值爲1的信號量 #define DECLARE_MUTEX(name) __DECLARE_SEMAPHORE_GENERIC(name,1) //聲明初始值爲0的信號量 #define DECLARE_MUTEX_LOCKED(name) __DECLARE_SEMAPHORE_GENERIC(name,0)
- down()
static inline void down(struct semaphore * sem) { #if WAITQUEUE_DEBUG CHECK_MAGIC(sem->__magic); #endif __asm__ __volatile__( "# atomic down operation\n\t" //鎖總線,對count減1 LOCK "decl %0\n\t" /* --sem->count */ "js 2f\n" "1:\n"//此時count大於等於0,返回down(),進入臨界區 ".section .text.lock,\"ax\"\n" "2:\tcall __down_failed\n\t"//此時count小於0,調用__down_failed "jmp 1b\n" ".previous" :"=m" (sem->count) :"c" (sem) :"memory"); }
- __down_failed()中調用了__down()
void __down(struct semaphore * sem) { struct task_struct *tsk = current; DECLARE_WAITQUEUE(wait, tsk); tsk->state = TASK_UNINTERRUPTIBLE; //將當前進程的等待隊列元素wait,鏈入隊列頭sem->wait的等待隊列的尾部 add_wait_queue_exclusive(&sem->wait, &wait); spin_lock_irq(&semaphore_lock); sem->sleepers++;//將等待進入臨界區的進程數加1 for (;;) { int sleepers = sem->sleepers; /* * Add "everybody else" into it. They aren't * playing, because we own the spinlock. */ //執行__down()函數的進程是因爲沒有進入臨界區,但此時可能有進程已經執行了up(),所以有必要再一次檢查count,避免無謂的等待進入睡眠而浪費資源 //atomic_add_negative()函數中執行sleepers-1加sem->count //若結果爲負數,返回非零,表示進程需要繼續等待 //若結果不爲負數,返回零,表示不需要等待,可以進入臨界區 if (!atomic_add_negative(sleepers - 1, &sem->count)) { sem->sleepers = 0;//設置等待進程數爲0 break;//跳出循環 } sem->sleepers = 1; /* us - see -1 above *///設置等待進程數爲1,它在這裏只表示有無進程需要等待,而不表示有多少進程需要等待 spin_unlock_irq(&semaphore_lock); schedule();//準備將此進程調度爲深度睡眠,即不會因爲信號而喚醒 tsk->state = TASK_UNINTERRUPTIBLE; spin_lock_irq(&semaphore_lock); } spin_unlock_irq(&semaphore_lock); remove_wait_queue(&sem->wait, &wait);//將此進程移出等待隊列 tsk->state = TASK_RUNNING;//設置此進程爲運行狀態 wake_up(&sem->wait);//返回之前喚醒等待隊列中的其他進程 }
up()(V操作)
- up()
static inline void up(struct semaphore * sem) { #if WAITQUEUE_DEBUG CHECK_MAGIC(sem->__magic); #endif __asm__ __volatile__( "# atomic up operation\n\t" //鎖總線,對count加1,這和前面的atomic_add_negative()函數的作用又對起來了 LOCK "incl %0\n\t" /* ++sem->count */ "jle 2f\n" "1:\n" ".section .text.lock,\"ax\"\n" "2:\tcall __up_wakeup\n\t"//當count小於等於0時,調用__up_wakeup() "jmp 1b\n" ".previous" :"=m" (sem->count) :"c" (sem) :"memory"); }
- __up_wakeup()中調用了__up(),__up()中調用了wake_up()
//wake_up()是宏函數,其中調用了__wake_up()函數 #define wake_up(x) __wake_up((x),TASK_UNINTERRUPTIBLE | TASK_INTERRUPTIBLE,WQ_FLAG_EXCLUSIVE)
- __wake_up()
//其中調用了__wake_up_common(),注意最後一個參數傳的是0 void __wake_up(wait_queue_head_t *q, unsigned int mode, unsigned int wq_mode) { __wake_up_common(q, mode, wq_mode, 0); }
- __wake_up_common()
static inline void __wake_up_common (wait_queue_head_t *q, unsigned int mode, unsigned int wq_mode, const int sync) { struct list_head *tmp, *head; struct task_struct *p, *best_exclusive; unsigned long flags; int best_cpu, irq; if (!q) goto out; best_cpu = smp_processor_id(); irq = in_interrupt(); best_exclusive = NULL; wq_write_lock_irqsave(&q->lock, flags); head = &q->task_list; tmp = head->next; while (tmp != head) { unsigned int state; wait_queue_t *curr = list_entry(tmp, wait_queue_t, task_list); tmp = tmp->next; p = curr->task; state = p->state; if (state & mode) { /* * If waking up from an interrupt context then * prefer processes which are affine to this * CPU. */ //此函數的作用就是遍歷等待隊列,依次喚醒符合條件的進程,如果喚醒的進程TASK_EXCLUSIVE爲1,就停止喚醒其餘進程,被喚醒的進程在__down()中繼續執行 if (irq && (curr->flags & wq_mode & WQ_FLAG_EXCLUSIVE)) { if (!best_exclusive) best_exclusive = p; if (p->processor == best_cpu) { best_exclusive = p; break; } } else { if (sync) wake_up_process_synchronous(p); else wake_up_process(p); if (curr->flags & wq_mode & WQ_FLAG_EXCLUSIVE) break; } } } if (best_exclusive) { if (sync) wake_up_process_synchronous(best_exclusive); else wake_up_process(best_exclusive); } wq_write_unlock_irqrestore(&q->lock, flags); out: return; }
*先去吃個飯。。一會來繼續寫spin_lock()、spin_unlock()。。→_→*