Re: [PATCH v5 2/5] serial: serialize asynchronous write wakeups with the port lock

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] serial: clps711x: deadlock between uart_clps711x_int_tx() and stop_tx()
--

commit dc0a2d4e309be9916a0d844a2a8caf7397942a8f
Author: James Hilliard <james.hilliard1@gmail.com>

serial: serialize asynchronous write wakeups with the port lock

This commit acquires the UART port lock around asynchronous write wakeups
and transmit helpers in several serial drivers so that uart_write_wakeup()
is serialized with port-locked serial-core state changes.

> diff --git a/drivers/tty/serial/clps711x.c b/drivers/tty/serial/clps711x.c
> index 83186bf500027..054b9d8359de2 100644
> --- a/drivers/tty/serial/clps711x.c
> +++ b/drivers/tty/serial/clps711x.c
> @@ -174,8 +174,10 @@ static irqreturn_t uart_clps711x_int_tx(int irq, void *dev_id)
>  			break;
>  	}
>  
> -	if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
> -		uart_write_wakeup(port);
> +	scoped_guard(uart_port_lock_irqsave, port) {
> +		if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
> +			uart_write_wakeup(port);
> +	}

[Severity: High]
Can acquiring port->lock in uart_clps711x_int_tx() deadlock with
uart_clps711x_stop_tx()?

The serial core invokes uart_ops->stop_tx() (uart_clps711x_stop_tx()) with
port->lock held, for example during uart_stop(), uart_flush_buffer(), or
uart_port_shutdown(). Inside uart_clps711x_stop_tx(), it calls synchronous
disable_irq(port->irq) while holding port->lock:

drivers/tty/serial/clps711x.c:uart_clps711x_stop_tx() {
	struct clps711x_port *s = dev_get_drvdata(port->dev);

	if (s->tx_enabled) {
		disable_irq(port->irq);
		s->tx_enabled = 0;
	}
}

If uart_clps711x_int_tx() is running concurrently (on another CPU or in a
threaded interrupt handler) and attempts to acquire port->lock while
uart_clps711x_stop_tx() holds port->lock and waits inside
disable_irq(port->irq) for the interrupt handler to complete, won't both
contexts deadlock waiting on each other?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261024-submit-serial-8250-dma-cancel-v1-v5-0-9e52968928d2@gmail.com?part=2