uart驱动与tty驱动

一:前言 接着前面的终端控制台分析,接下来分析serial的驱动.在linux中,serial也对应着终端,通常被称为串口终端. 在shell上,我们看到的/dev/ttyS*就是串口终端所对应的设备节点.在分析具体的serial驱动之前.有必要先 分析uart驱动架构.uart是Universal Asynchronous Receiver and Transmitter的缩写.翻译成中文 即为”通用异步收发器”.它是串口设备驱动的封装层. 二:uart驱动架构概貌 如下图所示: 上图中红色部份标识即为uart部份的操作. 从上图可以看到,uart设备是继tty_driver的又一层封装.实际上uart_driver就是对应tty_driver. 在它的操作函数中,将操作转入uart_port.在写操作的时候,先将数据放入一个叫做circ_buf的环形缓存 区.然后uart_port从缓存区中取数据,将其写入到串口设备中.当uart_port从serial设备接收到数据时 ,会将设备放入对应line discipline的缓存区中.这样.用户在编写串口驱动的时候,只先要注册一个 uart_driver.它的主要作用是定义设备节点号.然后将对设备的各项操作封装在uart_port.驱动工程师没 必要关心上层的流程,只需按硬件规范将uart_port中的接口函数完成就可以了. 三:uart驱动中重要的数据结构及其关联 我们可以自己考虑下,基于上面的架构代码应该要怎么写.首先考虑以下几点: 1: 一个uart_driver通常会注册一段设备号.即在用户空间会看到uart_driver对应有多个设备节点.例如: /dev/ttyS0  /dev/ttyS1 每个设备节点是对应一个具体硬件的,从上面的架构来看,每个设备文件应该对应一个uart_port. 也就是说:uart_device怎么同多个uart_port关系起来?怎么去区分操作的是哪一个设备文件? 2:每个uart_port对应一个circ_buf,所以uart_port必须要和这个缓存区关系起来 回忆tty驱动架构中.tty_driver有一个叫成员指向一个数组,即tty->ttys.每个设备文件对应设数组中的 一项.而这个数组所代码的数据结构为tty_struct. 相应的tty_struct会将tty_driver和ldisc关联起来. 那在uart驱动中,是否也可用相同的方式来处理呢? 将uart驱动常用的数据结构表示如下: 结合上面提出的疑问.可以很清楚的看懂这些结构的设计. 四:uart_driver的注册操作 Uart_driver注册对应的函数为: uart_register_driver()代码如下: int uart_register_driver(struct uart_driver *drv) {      struct tty_driver *normal = NULL;      int i, retval;      BUG_ON(drv->state);      /*       * Maybe we should be using a slab cache for this, especially if       * we have a large number of ports to handle.       */      drv->state = kzalloc(sizeof(struct uart_state) * drv->nr, GFP_KERNEL);      retval = -ENOMEM;      if (!drv->state)          goto out;      normal  = alloc_tty_driver(drv->nr);      if (!normal)          goto out;      drv->tty_driver = normal;      normal->owner      = drv->owner;      normal->driver_name    = drv->driver_name;      normal->name       = drv->dev_name;      normal->major      = drv->major;      normal->minor_start    = drv->minor;      normal->type       = TTY_DRIVER_TYPE_SERIAL;      normal->subtype        = SERIAL_TYPE_NORMAL;      normal->init_termios   = tty_std_termios;      normal->init_termios.c_cflag = B9600 | CS8 | CREAD | HUPCL | CLOCAL;      normal->init_termios.c_ispeed = normal->init_termios.c_ospeed = 9600;      normal->flags      = TTY_DRIVER_REAL_RAW | TTY_DRIVER_DYNAMIC_DEV;      normal->driver_state    = drv;      tty_set_operations(normal, &uart_ops);      /*       * Initialise the UART state(s).       */      for (i = 0; i nr; i++) {          struct uart_state *state = drv->state + i;          state->close_delay     = 500;    /* .5 seconds */          state->closing_wait    = 30000;  /* 30 seconds */          mutex_init(&state->mutex);      }      retval = tty_register_driver(normal); out:      if (retval          put_tty_driver(normal);          kfree(drv->state);      }      return retval; } 从上面代码可以看出.uart_driver中很多数据结构其实就是tty_driver中的.将数据转换为tty_driver 之后,注册tty_driver.然后初始化uart_driver->state的存储空间.这样,就会注册uart_driver->nr 个设备节点.主设备号为uart_driver-> major. 开始的次设备号为uart_driver-> minor.值得注意的 是.在这里将tty_driver的操作集统一设为了uart_ops.其次,在tty_driver-> driver_state保存了这 个uart_driver.这样做是为了在用户空间对设备文件的操作时,很容易转到对应的uart_driver.另外: tty_driver的flags成员值为: TTY_DRIVER_REAL_RAW | TTY_DRIVER_DYNAMIC_DEV.里面包含有 TY_DRIVER_DYNAMIC_DEV标志.结合之前对tty的分析.如果包含有这个标志,是不会在初始化的时候去注册 device.也就是说在/dev/下没有动态生成结点(如果是/dev下静态创建了这个结点就另当别论了^_^). 流程图如下: 五: uart_add_one_port()操作 在前面提到.在对uart设备文件过程中.会将操作转换到对应的port上,这个port跟uart_driver是怎么 关联起来的呢?这就是uart_add_ont_port()的主要工作了.顾名思义,这个函数是在uart_driver增加 一个port.代码如下: int uart_add_one_port(struct uart_driver *drv, struct uart_port *port) {      struct uart_state *state;      int ret = 0;      struct device *tty_dev;      BUG_ON(in_interrupt());      if (port->line >= drv->nr)          return -EINVAL;      state = drv->state + port->line;      mutex_lock(&port_mutex);      mutex_lock(&state->mutex);      if (state->port) {          ret = -EINVAL;          goto out;      }      state->port = port;      state->pm_state = -1;      port->cons = drv->cons;      port->info = state->info;      /*       * If this port is a console, then the spinlock is already       * initialised.       */      if (!(uart_console(port) && (port->cons->flags & CON_ENABLED))) {          spin_lock_init(&port->lock);          lockdep_set_class(&port->lock, &port_lock_key);      }      uart_configure_port(drv, state, port);      /*       * Register the port whether it's detected or not.  This allows       * setserial to be used to alter this ports parameters.       */      tty_dev = tty_register_device(drv->tty_driver, port->line, port->dev);      if (likely(!IS_ERR(tty_dev))) {          device_can_wakeup(tty_dev) = 1;          device_set_wakeup_enable(tty_dev, 0);      } else          printk(KERN_ERR "Cannot register tty device on line %d\n",                 port->line);      /*       * Ensure UPF_DEAD is not set.       */      port->flags &= ~UPF_DEAD; out:      mutex_unlock(&state->mutex);      mutex_unlock(&port_mutex);      return ret; } 首先这个函数不能在中断环境中使用. Uart_port->line就是对uart设备文件序号.它对应的也就是 uart_driver->state数组中的uart_port->line项.它主要初始化对应uart_driver->state项. 接着调用uart_configure_port()进行port的自动配置.然后注册tty_device.如果用户空间运行了 udev或者已经配置好了hotplug.就会在/dev下自动生成设备文件了. 操作流程图如下所示: 六:设备节点的open操作 在用户空间执行open操作的时候,就会执行uart_ops->open. Uart_ops的定义如下: static const struct tty_operations uart_ops = {      .open         = uart_open,      .close        = uart_close,      .write        = uart_write,      .put_char = uart_put_char,      .flush_chars  = uart_flush_chars,      .write_room   = uart_write_room,      .chars_in_buffer= uart_chars_in_buffer,      .flush_buffer = uart_flush_buffer,      .ioctl        = uart_ioctl,      .throttle = uart_throttle,      .unthrottle   = uart_unthrottle,      .send_xchar   = uart_send_xchar,      .set_termios  = uart_set_termios,      .stop         = uart_stop,      .start        = uart_start,      .hangup       = uart_hangup,      .break_ctl    = uart_break_ctl,      .wait_until_sent= uart_wait_until_sent, #ifdef CONFIG_PROC_FS      .read_proc    = uart_read_proc, #endif      .tiocmget = uart_tiocmget,      .tiocmset = uart_tiocmset, }; 对应open的操作接口为uart_open.代码如下: static int uart_open(struct tty_struct *tty, struct file *filp) {      struct uart_driver *drv = (struct uart_driver *)tty->driver->driver_state;      struct uart_state *state;      int retval, line = tty->index;      BUG_ON(!kernel_locked());      pr_debug("uart_open(%d) called\n", line);      /*       * tty->driver->num won't change, so we won't fail here with       * tty->driver_data set to something non-NULL (and therefore       * we won't get caught by uart_close()).       */      retval = -ENODEV;      if (line >= tty->driver->num)          goto fail;      /*       * We take the semaphore inside uart_get to guarantee that we won't       * be re-entered while allocating the info structure, or while we       * request any IRQs that the driver may need.  This also has the nice       * side-effect that it delays the action of uart_hangup, so we can       * guarantee that info->tty will always contain something reasonable.       */      state = uart_get(drv, line);      if (IS_ERR(state)) {          retval = PTR_ERR(state);          goto fail;      }      /*       * Once we set tty->driver_data here, we are guaranteed that       * uart_close() will decrement the driver module use count.       * Any failures from here onwards should not touch the count.       */      tty->driver_data = state;      tty->low_latency = (state->port->flags & UPF_LOW_LATENCY) ? 1 : 0;      tty->alt_speed = 0;      state->info->tty = tty;      /*       * If the port is in the middle of closing, bail out now.       */      if (tty_hung_up_p(filp)) {          retval = -EAGAIN;          state->count--;          mutex_unlock(&state->mutex);          goto fail;      }      /*       * Make sure the device is in D0 state.       */      if (state->count == 1)          uart_change_pm(state, 0);      /*       * Start up the serial port.       */      retval = uart_startup(state, 0);      /*       * If we succeeded, wait until the port is ready.       */      if (retval == 0)          retval = uart_block_til_ready(filp, state);      mutex_unlock(&state->mutex);      /*       * If this is the first open to succeed, adjust things to suit.       */      if (retval == 0 && !(state->info->flags & UIF_NORMAL_ACTIVE)) {          state->info->flags |= UIF_NORMAL_ACTIVE;          uart_update_termios(state);      } fail:      return retval; } 在这里函数里,继续完成操作的设备文件所对应state初始化.现在用户空间open这个设备了.即要对这个 文件进行操作了.那uart_port也要开始工作了.即调用uart_startup()使其进入工作状态.当然,也需 要初始化uart_port所对应的环形缓冲区circ_buf.即state->info-> xmit.特别要注意,在这里将 tty->driver_data = state;这是因为以后的操作只有port相关了,不需要去了解uart_driver的 相关信息.跟踪看一下里面调用的两个重要的子函数. uart_get()和uart_startup().先分析 uart_get().代码如下: static struct uart_state *uart_get(struct uart_driver *drv, int line) {      struct uart_state *state;      int ret = 0;      state = drv->state + line;      if (mutex_lock_interruptible(&state->mutex)) {          ret = -ERESTARTSYS;          goto err;      }      state->count++;      if (!state->port || state->port->flags & UPF_DEAD) {          ret = -ENXIO;          goto err_unlock;      }      if (!state->info) {          state->info = kzalloc(sizeof(struct uart_info), GFP_KERNEL);          if (state->info) {               init_waitqueue_head(&state->info->open_wait);               init_waitqueue_head(&state->info->delta_msr_wait);               /*                * Link the info into the other structures.                */               state->port->info = state->info;               tasklet_init(&state->info->tlet, uart_tasklet_action,                         (unsigned long)state);          } else {               ret = -ENOMEM;               goto err_unlock;          }      }      return state; err_unlock:      state->count--;      mutex_unlock(&state->mutex); err:      return ERR_PTR(ret); } 从代码中可以看出.这里注要是操作是初始化state->info.注意port->info就是state->info的一个 副本.即port直接通过port->info可以找到它要操作的缓存区. uart_startup()代码如下: static int uart_startup(struct uart_state *state, int init_hw) {      struct uart_info *info = state->info;      struct uart_port *port = state->port;      unsigned long page;      int retval = 0;      if (info->flags & UIF_INITIALIZED)          return 0;      /*       * Set the TTY IO error marker - we will only clear this       * once we have successfully opened the port.  Also set       * up the tty->alt_speed kludge       */      set_bit(TTY_IO_ERROR, &info->tty->flags);      if (port->type == PORT_UNKNOWN)          return 0;      /*       * Initialise and allocate the transmit and temporary       * buffer.       */      if (!info->xmit.buf) {          page = get_zeroed_page(GFP_KERNEL);          if (!page)               return -ENOMEM;          info->xmit.buf = (unsigned char *) page;          uart_circ_clear(&info->xmit);      }      retval = port->ops->startup(port);      if (retval == 0) {          if (init_hw) {               /*                * Initialise the hardware port settings.                */               uart_change_speed(state, NULL);               /*                * Setup the RTS and DTR signals once the                * port is open and ready to respond.                */               if (info->tty->termios->c_cflag & CBAUD)                    uart_set_mctrl(port, TIOCM_RTS | TIOCM_DTR);          }          if (info->flags & UIF_CTS_FLOW) {               spin_lock_irq(&port->lock);               if (!(port->ops->get_mctrl(port) & TIOCM_CTS))                    info->tty->hw_stopped = 1;               spin_unlock_irq(&port->lock);          }          info->flags |= UIF_INITIALIZED;          clear_bit(TTY_IO_ERROR, &info->tty->flags);      }      if (retval && capable(CAP_SYS_ADMIN))          retval = 0;      return retval; } 在这里,注要完成对环形缓冲,即info->xmit的初始化.然后调用port->ops->startup( )将这个port带 入到工作状态.其它的是一个可调参数的设置,就不详细讲解了. 七:设备节点的write操作 Write操作对应的操作接口为uart_write( ).代码如下: static int uart_write(struct tty_struct *tty, const unsigned char *buf, int count) {      struct uart_state *state = tty->driver_data;      struct uart_port *port;      struct circ_buf *circ;      unsigned long flags;      int c, ret = 0;      /*       * This means you called this function _after_ the port was       * closed.  No cookie for you.       */      if (!state || !state->info) {          WARN_ON(1);          return -EL3HLT;      }      port = state->port;      circ = &state->info->xmit;      if (!circ->buf)          return 0;      spin_lock_irqsave(&port->lock, flags);      while (1) {          c = CIRC_SPACE_TO_END(circ->head, circ->tail, UART_XMIT_SIZE);          if (count               c = count;          if (c               break;          memcpy(circ->buf + circ->head, buf, c);          circ->head = (circ->head + c) & (UART_XMIT_SIZE - 1);          buf += c;          count -= c;          ret += c;      }      spin_unlock_irqrestore(&port->lock, flags);      uart_start(tty);      return ret; } Uart_start()代码如下: static void uart_start(struct tty_struct *tty) {      struct uart_state *state = tty->driver_data;      struct uart_port *port = state->port;      unsigned long flags;      spin_lock_irqsave(&port->lock, flags);      __uart_start(tty);      spin_unlock_irqrestore(&port->lock, flags); } static void __uart_start(struct tty_struct *tty) {      struct uart_state *state = tty->driver_data;      struct uart_port *port = state->port;      if (!uart_circ_empty(&state->info->xmit) && state->info->xmit.buf &&          !tty->stopped && !tty->hw_stopped)          port->ops->start_tx(port); } 显然,对于write操作而言,它就是将数据copy到环形缓存区.然后调用port->ops->start_tx()将数据写 到硬件寄存器. 八:Read操作 Uart的read操作同Tty的read操作相同,即都是调用ldsic->read()读取read_buf中的内容.有对这部份 内容不太清楚的,参阅http://blog.youkuaiyun.com/xp_super/article/details/8011062     九:小结 本小节是分析serial驱动的基础.在理解了tty驱动架构之后,再来理解uart驱动架构应该不是很难.随着我们 在linux设备驱动分析的深入,越来越深刻的体会到,linux的设备驱动架构很多都是相通的.只要深刻理解了一 种驱动架构.举一反三.也就很容易分析出其它架构的驱动了.