Hi!

> [ Upstream commit ba3684f99f1b25d2a30b6956d02d339d7acb9799 ]
> 
> The function msm_wait_for_xmitr can be taken with interrupts
> disabled. In order to avoid a potential system lockup - demonstrated
> under stress testing conditions on SoC QCS404/5 - make sure we wait
> for a bounded amount of time.
> 
> Tested on SoC QCS404.

How long did it take to timeout?

Because... this is supposed to loop for 0.5 second with interrupts
disabled, but 500000*udelay(1) is probably going to wait for more than
that.

Is 500msec reasonable with interrupts disabled?

Should it use something like 5000*udelay(100), instead, as that has
chance to result in closer-to-500msec wait?

> +++ b/drivers/tty/serial/msm_serial.c
> @@ -383,10 +383,14 @@ static void msm_request_rx_dma(struct msm_port *msm_port, resource_size_t base)
>  
>  static inline void msm_wait_for_xmitr(struct uart_port *port)
>  {
> +	unsigned int timeout = 500000;
> +
>  	while (!(msm_read(port, UART_SR) & UART_SR_TX_EMPTY)) {
>  		if (msm_read(port, UART_ISR) & UART_ISR_TX_READY)
>  			break;
>  		udelay(1);
> +		if (!timeout--)
> +			break;
>  	}
>  	msm_write(port, UART_CR_CMD_RESET_TX_READY, UART_CR);
>  }

Plus, should it do some kind of dev_err() to let users know that
something went very wrong with their serial?

Thanks,
								Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html