1]有关写操作:
1.1如果上层使用poll系统调用监测写事件,就是知道什么时候可以进行写操作,tty必须实现的接口
其中有两个tty_ops write相关的函数,其中tty_chars_in_buffer是必须实现的unsigned int n_tty_poll(struct tty_struct *tty, struct file *file, poll_table *wait)
{
if (tty->ops->write && !tty_is_writelocked(tty) &&
tty_chars_in_buffer(tty) < WAKEUP_CHARS &&
tty_write_room(tty) > 0)
mask |= POLLOUT | POLLWRNORM;
return mask;
}
1.1.1 tty_chars_in_buffer - characters pending, *Return the number of bytes of data in the device output queue.
int tty_chars_in_buffer(struct tty_struct *tty){
if (tty->ops->chars_in_buffer)
return tty->ops->chars_in_buffer(tty);
else
return 0;
}
1.1.2 tty_write_room:没有该函数,就没有流控,就可能丢失数据
* tty_write_room - write queue space* @tty: terminal
*
* Return the number of bytes that can be queued to this device
* at the present time. The result should be treated as a guarantee
* and the driver cannot offer a value it later shrinks by more than
* the number of bytes written. If no method is provided 2K is always
* returned and data may be lost as there will be no flow control.
int tty_write_room(struct tty_struct *tty)
{
if (tty->ops->write_room)
return tty->ops->write_room(tty);
return 2048;
}
1.2 使用流控函数tty_throttle
这个函数只在读操作中使用,不用在writ 操作中。总结:设想这样一种情形:用户要写一段数据。
如果使用poll监测是否可写,如果可写则写入数据,这种方法有有个地方唤醒
tty->write_wait,或者poll的参数带个timeout,没timeout就调用n_tty_poll查询一次write_room的状态。
如果不使用poll函数,直接写,如果底层函数在虽然请求为n个字节的大写,却返回为0,则睡眠当前进程,等待 tty->write_wait;
如果底层函数在虽然请求为n个字节的大写,不返回为0,直到写完数据才返回,则不会睡眠当前进程。
2]有关读操作
2.1如果上层使用poll系统调用监测读事件,就是知道什么时候可以进行读操作,tty必须实现的接口
static unsigned int n_tty_poll(struct tty_struct *tty, struct file *file, poll_table *wait){
if (input_available_p(tty, TIME_CHAR(tty) ? 0 : MIN_CHAR(tty)))
mask |= POLLIN | POLLRDNORM;
return mask;
}
2.1.1 input_available_p
该函数调用tty_flush_to_ldisc触发一次tty_buff到tty_read buffer的一次copy,最后判断read_cnt有多少数据需要读出static inline int input_available_p(struct tty_struct *tty, int amt)
{
tty_flush_to_ldisc(tty);
if (tty->icanon && !L_EXTPROC(tty)) {
if (tty->canon_data)
return 1;
} else if (tty->read_cnt >= (amt ? amt : 1))
return 1;
return 0;
}
2.2 使用流控函数tty_throttle
2.2.1 在数据从tty_buffer到tty_read buffer的work中调用tty_throttle
flush_to_ldisc -> n_tty_receive_buf ->{* Check the remaining room for the input canonicalization
* mode. We don't want to throttle the driver if we're in
* canonical mode and don't have a newline yet!
if (tty->receive_room < TTY_THRESHOLD_THROTTLE)
tty_throttle(tty);
}
void tty_throttle(struct tty_struct *tty)
{
mutex_lock(&tty->termios_mutex);
/* check TTY_THROTTLED first so it indicates our state */
if (!test_and_set_bit(TTY_THROTTLED, &tty->flags) &&
tty->ops->throttle)
tty->ops->throttle(tty);
mutex_unlock(&tty->termios_mutex);
}
2.2.2 n_tty_read 调用tty_unthrottle
n_tty_read->{/* If there is enough space in the read buffer now, let the
* low-level driver know. We use n_tty_chars_in_buffer() to
* check the buffer, as it now knows about canonical mode.
* Otherwise, if the driver is throttled and the line is
* longer than TTY_THRESHOLD_UNTHROTTLE in canonical mode,
* we won't get any more characters.
*/
if (n_tty_chars_in_buffer(tty) <= TTY_THRESHOLD_UNTHROTTLE) {
n_tty_set_room(tty);
check_unthrottle(tty);
}
}
static void check_unthrottle(struct tty_struct *tty)
{
if (tty->count)
tty_unthrottle(tty);
}
何时tty->count变化啊?tty_standard_install and tty_reopen
void tty_unthrottle(struct tty_struct *tty)
{
mutex_lock(&tty->termios_mutex);
if (test_and_clear_bit(TTY_THROTTLED, &tty->flags) &&
tty->ops->unthrottle)
tty->ops->unthrottle(tty);
mutex_unlock(&tty->termios_mutex);
}
2.2.3 如果没有实现流控函数 tty_un/throttle,结果会是怎样?
只要有数据产生,n_tty_receive_buf函数就会把数据从tty_buf拷贝到tty read_buffer这个循环buffer中,这就会造成数据的丢失。所以数据要及时读出,或者增大read buffer的大小。
总结:设想这样一种情形:用户要读一段数据。
如果使用poll监测是否有可读的数据,如果可有就调用读函数。读操作中有多处调用唤醒tty->read_wait可以唤醒poll。
如果不使用poll函数,直接读,如果没有数据可读则睡眠当前进程,等待 tty->read_wait被唤醒;
**/